Assessing the Business Case for Integrated
Collision Avoidance Systems on Transit Buses
August 2007
REPORT DOCUMENTATION PAGE
Form Approved
OMB No. 0704- 0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202- 4302, and to the Office of Management and Budget, Paperwork Reduction Project ( 0704- 0188), Washington, DC 20503.
1. AGENCY USE ONLY ( Leave blank)
2. REPORT DATE
August 2007
3. REPORT TYPE AND DATES COVERED
4. TITLE AND SUBTITLE
Assessing the Business Case for Integrated Collision Avoidance Systems on Transit Buses
6. AUTHOR( S)
Travis Dunn, Richard Laver, Douglas Skorupski, Deborah Zyrowski
5. FUNDING NUMBERS
7. PERFORMING ORGANIZATION NAME( S) AND ADDRESS( ES)
Booz Allen Hamilton, Inc.
8283 Greensboro Drive
McLean, VA 22102
8. PERFORMING ORGANIZATION REPORT NUMBER
9. SPONSORING/ MONITORING AGENCY NAME( S) AND ADDRESS( ES)
Federal Transit Administration
U. S. Department of Transportation
Washington, DC 20590
10. SPONSORING/ MONITORING AGENCY REPORT NUMBER
11. Supplementary Notes.
12a. DISTRIBUTION/ AVAILABILITY STATEMENT
Available From: National Technical Information Service/ NTIS, 5285 Port Royal Road, Springfield, Virginia 22161. Phone 703.605.6000, Fax 703.605.6900, Email [ orders@ ntis. fedworld. gov]
12b. DISTRIBUTION CODE
13. ABSTRACT ( Maximum 200 words)
This document presents an analysis of Integrated Vehicle Based Safety Systems ( IVBSS) for transit buses. The study took a three- pronged approach. The first was an analysis of the available IVBSS products, possible future products and the technologies. The second was a benefit- cost analysis of transit IVBSS. The third assessed the receptiveness among transit operators to use IVBSS products and the willingness of manufacturers to develop them.
This study used the National Transit Database and crash data from 6 U. S. transit operators. The data show that there is an average of 1.5 collisions per transit bus and related annual costs of over $ 4,000. Of the technologies evaluated, only side object detection systems showed the potential to be cost effective. In general, transit agencies are receptive to in- vehicle safety devices when there is evidence of their effectiveness. Several vendors currently offer products while others are awaiting commitments from the U. S. DOT or coordinated transit industry interest before developing their products. It is recommended that the U. S. DOT pursue operational tests of the side object detection system and other stronger- performing systems in order to validate the findings of this study.
15. NUMBER OF PAGES
14. SUBJECT TERMS
Integrated Vehicle Based Safety Systems, Integrated Collision Warning System, Forward Collision Warning System, Side Collision Warning System, Side Object Detection System, Pedestrian Detection, Lane Departure Warning, Rear Object Detection System, Rear Collision Warning System, Transit Bus, Transit Safety
16. PRICE CODE
17. SECURITY CLASSIFICATION OF REPORT
Unclassified
18. SECURITY CLASSIFICATION OF THIS PAGE
Unclassified
19. SECURITY CLASSIFICATION OF ABSTRACT
Unclassified
20. LIMITATION OF ABSTRACT
NSN 7540- 01- 280- 5500 Standard Form 298 ( Rev. 2- 89)
Prescribed by ANSI Std. 239- 18298- 102 DISCLAIMER NOTICE
This document is disseminated under the sponsorship of the U. S. Department of Transportation in the interest of information exchange. The United States Government assumes no liability for its contents or use thereof.
The United States Government does not endorse products of manufacturers. Trade or manufacturers’ names appear herein solely because they are considered essential to the objective of this report.
Assessing the Business Case for Integrated Collision Avoidance Systems on Transit Buses
Final Report
August 2007
Prepared by:
Booz Allen Hamilton
8283 Greensboro Drive
McLean, VA 22102
Prepared for:
Office of Research, Demonstration and Innovation
Federal Transit Administration
1200 New Jersey Ave SE
Washington, DC 20590
And
ITS Joint Program Office
Federal Highway Administration
1200 New Jersey Ave SE
Washington, DC 20590 PREFACE
The work presented in this report was conducted by Booz Allen Hamilton. This study was sponsored by the Federal Transit Administration ( FTA) and the ITS Joint Program Office at the U. S. Department of Transportation ( U. S. DOT) as part of the Integrated Vehicle- Based Safety Systems ( IVBSS) program. Results generated from this study will be an important reference for the transit IVBSS project teams with their ongoing work of studying and developing transit IVBSS technologies.
The authors would like to acknowledge Mr. Sébastien Renaud of FTA and Eric Traube, Kathryn Wochinger, and David Yang of Noblis for their contributions to this report. Assessing the Business Case for Transit IVBSS Final Report
Table of Contents
Table of Contents....................................................................................................................... ................... i
List of Figures........................................................................................................................ ...................... ii
List of Tables......................................................................................................................... ..................... iii
List of Acronyms....................................................................................................................... ................. iv
1.0 Executive Summary........................................................................................................................ 1
1.1 Background..................................................................................................................... ............ 1
1.2 Study Objectives and Approach.................................................................................................. 1
1.3 Key Findings....................................................................................................................... ....... 2
1.4 Recommendations................................................................................................................ ...... 4
2.0 Introduction................................................................................................................... .................. 7
2.1 Background..................................................................................................................... ............ 7
2.2 Study Objectives..................................................................................................................... .. 10
2.3 Study Approach....................................................................................................................... . 11
2.4 Report Format......................................................................................................................... .. 14
3.0 Technology Assessment................................................................................................................. 15
3.1 Collision Avoidance Systems.................................................................................................... 16
3.2 Sensor Technologies.................................................................................................................. 17
3.3 System Manufacturers............................................................................................................... 18
3.4 Prototype Systems..................................................................................................................... 26
4.0 Benefit- Cost Analysis.................................................................................................................... 34
4.1 Benefits of IVBSS Deployment ( step 1)................................................................................... 35
4.2 Costs of IVBSS Deployment ( step 2)........................................................................................ 49
4.3 Benefit- Cost Results ( step 3)..................................................................................................... 52
4.4 Sensitivity Analysis................................................................................................................... 57
4.5 Other Implications of This Analysis.......................................................................................... 61
5.0 Risk Assessment and Market Viability.......................................................................................... 65
5.1 Agency View........................................................................................................................... . 65
5.2 Vendor View........................................................................................................................... . 75
5.3 Market Viability...................................................................................................................... . 76
6.0 Next Steps: Operational Tests and Deployment Strategies............................................................ 86
6.1 Research and Testing Activities................................................................................................ 87
6.2 Communication and Outreach Efforts....................................................................................... 91
6.3 Programs for the Supplier Community...................................................................................... 91
6.4 Technology Monitoring............................................................................................................. 92
7.0 Findings and Recommendations.................................................................................................... 94
7.1 Key Findings....................................................................................................................... ..... 94
7.2 Key Recommendations.............................................................................................................. 98
Appendix A: IVBSS Outreach Session Attendees.................................................................................... 100
Appendix B: Agency Interview Guide...................................................................................................... 102
Appendix C: Technology Effectiveness Assessment................................................................................ 136
Appendix D: Sensitivity Analysis on Alternative Input Variables........................................................... 140
Appendix E: References..................................................................................................................... ..... 142 Federal Transit Administration i
Assessing the Business Case for Transit IVBSS Final Report
List of Figures
Figure 2- 1: Collision Countermeasure Strategies......................................................................................... 8
Figure 2- 2: Typology of In- Vehicle Safety Systems.................................................................................. 10
Figure 2- 3: U. S. DOT Decision- Making Process for Transit IVBSS......................................................... 11
Figure 3- 1: Diagram of Collision Avoidance Systems ( Plan View of a Bus Facing to the Right)............. 16
Figure 3- 2: Clever Device’s Ultrasonic Sensor.......................................................................................... 19
Figure 3- 3: Seymor System Sensor Locations............................................................................................ 20
Figure 3- 4: Seymor Modes of Operation.................................................................................................... 20
Figure 3- 5: Seymor Driver Visual Interface ( DVI)..................................................................................... 21
Figure 3- 6: Seymor DVI Locations............................................................................................................ 21
Figure 3- 7: EyeQ Chip and SeeQ Board..................................................................................................... 22
Figure 3- 8: Mobileye's Monocular System - Single Camera with Internal EyeQ Board............................ 23
Figure 3- 9: Mobileye's DVI Demonstrating FCW...................................................................................... 23
Figure 3- 10: Mobileye's DVI Demonstrating LDW................................................................................... 24
Figure 3- 11: Mobileye's DVI Demonstrating HMW.................................................................................. 24
Figure 3- 12: Pedestrian Detection............................................................................................................... 24
Figure 3- 13: RICWS Installed at AATA.................................................................................................... 26
Figure 3- 14: FCWS Sensor/ Camera Locations........................................................................................... 27
Figure 3- 15: FCWS DVI............................................................................................................................ 28
Figure 3- 16: SCWS Retractable Laser Scanner.......................................................................................... 29
Figure 3- 17: SamTrans and PAAC ICWS Buses........................................................................................ 30
Figure 3- 18: ICWS DVI............................................................................................................................ . 30
Figure 3- 19: Installed ICWS DVI............................................................................................................... 31
Figure 3- 20: WMATA Safety Warning Strobe Light................................................................................. 31
Figure 4- 1: IVBSS Benefit- Cost Computational Framework..................................................................... 35
Figure 4- 2: Collision Scenario Matrix........................................................................................................ 38
Figure 4- 3: Three Scenarios of Collision Type 1 (“ bus straight ahead – other vehicle from left”)............ 43
Figure 4- 4: Two Scenarios of Collision Type 4 (“ bus turning right, other vehicle from left”)....................... 46
Figure 4- 5: IVBSS Effectiveness in Preventing Frontal Collisions with Vehicles..................................... 47
Figure 4- 6: Collisions Prevented and Collision Costs by System............................................................... 55
Figure 4- 7: Benefit- Cost Ratio Variation with Collision Frequency for Standalone Systems................... 60
Figure 4- 8: Benefit- Cost Ratio Variation with Collision Frequency for Combination Systems................ 60
Figure 4- 9: Risk Assumption Profile for a Range of Insurance Arrangements.......................................... 62
Figure 4- 10: Conceptual Representation of the Relationship between Safety Record and Insurance Premiums....................................................................................................................... ............................ 64
Figure 5- 1: Coverage Zone of Two Alternative Sensor Arrangements...................................................... 71
Figure 5- 2: Existing Transit Vehicle Technologies.................................................................................... 74
Figure 5- 3: Percent of U. S. Transit Buses Exposed to Various Major and Non- Major Accident Rates per Bus per Year........................................................................................................................... ................... 78
Figure 5- 4: Benefit- Cost Ratios for Seven Standalone Systems and Bus Population, by Collision Involvement Rate........................................................................................................................... ............ 79
Figure 5- 5: Benefit- Cost Ratios for Technology Packages and Bus Population, by Collision Involvement Rate........................................................................................................................... ................................. 80
Figure 5- 6: Number of U. S. Transit Buses within Fleets of Various Sizes................................................ 82
Figure 6- 1: Road Map Steps....................................................................................................................... 86 Federal Transit Administration ii
Assessing the Business Case for Transit IVBSS Final Report
List of Tables
Table 1- 1: Range of Benefit- Cost Ratios for Standalone Safety Systems.................................................... 3
Table 1- 2: Range of Benefit- Cost Ratios for Integrated ( IVBSS) Systems.................................................. 3
Table 1- 3: Percent of Collisions Judged “ Avoidable” by Type of Collision................................................ 4
Table 2- 1: Safety Summary for Transit Buses vs. All Roadway Modes ( PMT= Passenger- Miles Traveled). 7
Table 2- 2: Mix of Collisions by Type for Transit Buses vs. Other Modes................................................... 7
Table 2- 3: Annual Frequency of Collisions by Type and Severity per 1000 Transit Buses....................... 13
Table 3- 1: Overview of Collision Avoidance Systems............................................................................... 15
Table 4- 1: Average Annual Number of Collisions per 1,000 Buses by Type and Severity....................... 39
Table 4- 2: Characterization of Six Participating Transit Agencies............................................................. 40
Table 4- 3: Variability in Mix of Collision Types Among Six Transit Agencies........................................ 40
Table 4- 4: Major Categories of Costs Related to Transit Bus Collisions................................................... 41
Table 4- 5: Annual Collision Cost Per Bus by Collision Type.................................................................... 42
Table 4- 6: Percent of Collisions Judged " Avoidable" by Type of Collision.............................................. 44
Table 4- 7 Scale Used to Rate Each System’s Effectiveness in Preventing Different Types of Collisions. 45
Table 4- 8: Sample Portion of the IVBSS Effectiveness Evaluation Framework ( from Appendix C)........ 46
Table 4- 9: Collision Prevention Rates by System and Collision Type....................................................... 47
Table 4- 10: Year 1 Benefits by System ( per equipped vehicle)................................................................. 49
Table 4- 11: Costs of Various Unbundled Safety Systems per Unit Assuming Fleet Size of 100............... 51
Table 4- 12: Costs of Various Integrated ( IVBSS) Safety Systems............................................................. 52
Table 4- 13: Benefit- Cost Ratio under Baseline Assumptions for Various Standalone Safety Systems..... 53
Table 4- 14: Benefit- Cost Ratio under Baseline Assumptions for Various Safety System Combinations.. 53
Table 4- 15: Payback Period under Baseline Assumptions for Various Standalone Safety Systems.......... 53
Table 4- 16: Payback Period under Baseline Assumptions for Various Safety System Combinations....... 54
Table 4- 17: Details about System Performance against Collision Types................................................... 55
Table 4- 18: Effectiveness Required to Break Even under Baseline Assumptions for Standalone Systems........................................................................................................................ ........................................... 56
Table 4- 19: Effectiveness Required to Break Even Under Baseline Assumptions for System Packages.. 57
Table 4- 20: Range of Collision Costs by Crash Type, Using 95% Confidence Interval on Cost Buildup. 58
Table 4- 21: Range of Unreported Collision Frequencies Per 1000 Vehicles, Using 95% Confidence Intervals...................................................................................................................... ............................... 58
Table 4- 22: Range of Benefit- Cost Ratios for Standalone Systems Based on 95% Confidence in Collision Costs and Frequencies.................................................................................................................... ............ 58
Table 4- 23: Range of Benefit- Cost Ratios for System Combinations Based on 95% Confidence in Collision Costs and Frequencies................................................................................................................. 59
Table 4- 24: Sensitivity Analysis Results For System Cost......................................................................... 61
Table 5- 1: Investment Priorities of Outreach Session Participants............................................................. 75
Table 5- 2: Number of Buses for Which Investment Is Financially Justifiable, by System........................ 81
Table 5- 3: Retrofit Market Sales Potential................................................................................................. 83
Table 5- 4: Market Sales Potential to Equip New Buses............................................................................. 84
Table 6- 1: Annual Collisions Per 1,000 Buses by Collision Type ( National Average).............................. 88
Table 7- 1: Range of Benefit- Cost Ratios for Standalone Safety Systems Using 95% Confidence Interval for Collision Cost and Frequency Inputs.................................................................................................... 95
Table 7- 2: Range of Benefit- Cost Ratios for System Combinations Using 95% Confidence Interval for Collision Cost and Frequency Inputs.......................................................................................................... 95
Table 7- 3: Effectiveness Required to Break Even under Baseline Assumptions for Standalone Safety Systems........................................................................................................................ .............................. 96
Table 7- 4: Effectiveness Required to Break Even Under Baseline Assumptions for System Packages.... 96 Federal Transit Administration iii
Assessing the Business Case for Transit IVBSS Final Report
List of Acronyms
AATA Ann Arbor Transportation Authority
AC Transit Alameda County Transit Authority
APC Automatic Passenger Counter
APTA American Public Transportation Association
AVL Automatic Vehicle Location
AWS Advance Warning System
BIFA Buses Involved in Fatal Accidents
CAD Computer- Aided Dispatch
Caltrans California Department of Transportation
CAS Collision Avoidance Systems
CTA Chicago Transit Authority
CWS Collision Warning Systems
CWS+ A system containing a FCWS, a LDWS and a PDS
DOT U. S. Department of Transportation
DVI Driver Visual Interface or Driver- Vehicle Interface
FARS Fatality Analysis Reporting System
FCSD Forward Collision/ Side Detection ( a combination of FCWS and SODS)
FCWS Forward Collision Warning System
FHWA Federal Highway Administration
FMCSA Federal Motor Carrier Safety Administration
FODS Forward Object Detection System
FTA Federal Transit Administration
GCRTA Greater Cleveland Regional Transit Authority
GES General Estimates System
GIS Geographical Information Systems
GM General Motors
GPS Global Positioning Systems
HMW Headway Monitoring and Warning
ICWS Integrated Collision Warning System
IT Information Technology
ITS Intelligent Transportation Systems
IVBSS Integrated Vehicle- Based Safety Systems
IVI Intelligent Vehicle Initiative
JPO Joint Program Office
LDWS Lane Departure Warning System
LED Light Emitting Diode
Federal Transit Administration iv
Assessing the Business Case for Transit IVBSS Final Report
LLS Laser Line Striper
Metro Los Angeles County Metropolitan Transportation Authority
MPH Miles per Hour
MTA Maryland Transit Administration
Muni San Francisco Municipal Railway
NCTD North County Transit District
NHTSA National Highway Transportation Safety Administration
NTD National Transit Database
O& M Operations and Maintenance
ODS Object Detection Systems
OEM Original Equipment Manufacturer
PAAC Port Authority of Allegheny County
Pace Pace Suburban Bus
PDO Property- Damage- Only
PDS Pedestrian Detection System
PED Pedestrian
PMT Passenger- Miles Traveled
RCWS Rear Collision Warning System
RICWS Rear Impact Collision Warning System
RODS Rear Object Detection System
RTD Regional Transportation District ( Denver)
SamTrans San Mateo County Transit District
SCWS Side Collision Warning System
SODS Side Object Detection System
TTC Time to Contact
UMTRI University of Michigan Transportation Research Institute
US United States
UTA Utah Transit Authority
VTA Santa Clara Valley Transportation Administration
WMATA Washington Metropolitan Area Transit Authority Federal Transit Administration v
Assessing the Business Case for Transit IVBSS Final Report
1.0 EXECUTIVE SUMMARY
1.1 Background
Transit buses are involved in approximately 100,000 collisions each year, leading to nearly 100 fatalities and 7,500 injuries. Transit bus operators must address all of the financial costs typically associated with these collisions, including damage repairs, claims payments, legal fees, workers’ compensation, and lost productivity. Moreover, public perceptions of safety can be tarnished by a single incident, eroding the trust and confidence of the public and generating unfavorable media attention. Collisions can also disrupt bus service and cause delays for all roadway users, inhibiting the operator’s ability to fulfill its mission to the public.
Transit operators continuously seek methods and products that reduce their exposure to safety hazards. However, investment specifically in advanced in- vehicle safety systems has been slow because of operator uncertainty about the effectiveness of such systems in preventing or mitigating collisions. Likewise, some suppliers of safety systems are reluctant to invest resources in developing transit- specific products while the operators’ potential demand remains uncertain.
In recognition of the potential to improve the performance and deployment rates of in- vehicle safety systems across all roadway travel modes, the U. S. Department of Transportation ( DOT) initiated the Integrated Vehicle- Based Safety Systems ( IVBSS) program. The specific goals of the IVBSS program are to integrate, simplify, and reduce the costs of safety technologies; increase safety benefits ( e. g., by reducing collision incident rates); improve overall safety system performance; improve acceptance of in- vehicle technologies; and enhance the marketability of safety devices. To date, the IVBSS program has included widespread demonstration and evaluation of safety systems designed specifically for passenger vehicles and heavy trucks, but not for transit vehicles.
1.2 Study Objectives and Approach
The purpose of this study is to evaluate the business case for ( or against) the development of integrated safety systems for transit buses. Specifically, this study addresses the question of whether the expected benefits from investing in these systems ( e. g., expected sales by vendors or reductions in accident costs to transit agencies) outweigh the costs. Based on the results of this study and other related studies, the U. S. DOT will determine whether bus transit safety systems warrant additional investment in operational tests, demonstrations, and evaluations. This report represents an input to that “ go/ no go” decision.
The business case evaluation presented here focuses on the following seven existing and potential safety systems, referred to collectively as “ collision avoidance systems:”
• Forward Collision Warning System ( FCWS)
• Rear Collision Warning System ( RCWS)
• Side Object Detection System ( SODS)
• Forward Object Detection System ( FODS)
• Rear Object Detection System ( RODS)
• Lane Departure Warning System ( LDWS)
• Pedestrian Detection System ( PDS)
Federal Transit Administration 1
Assessing the Business Case for Transit IVBSS Final Report
This study evaluated the technical, financial, and qualitative investment merits of these seven systems, both as standalone systems and as integrated ( i. e., IVBSS) investments, by following a three- pronged approach:
1. Technology Evaluation: Identified and evaluated the functional and technical characteristics of in- vehicle safety systems applicable to transit buses.
2. Benefit- Cost Analysis: Conducted a benefit- cost analysis of the safety systems ( both individually and in integrated “ IVBSS” packages) from the perspective of an investing transit agency ( to determine whether agency investment benefits exceed direct agency costs) and from the perspective of system vendors.
3. Industry Outreach: Conducted outreach sessions with transit operators to document their perceptions of these systems as well as the qualitative risks, rewards, and concerns likely to determine their interest in investing in these and similar systems.
1.3 Key Findings
Technology evaluation
The following are key findings regarding technology evaluation:
• Collision avoidance systems are divided into two categories: 1) object detection systems ( ODS) and 2) collision warning systems ( CWS). ODS monitor the area in close proximity to the vehicle and are designed to detect objects that are not within the view of the driver. Systems available at this time include forward, side, and rear object detection. Most manufacturers of non- video- based object detection systems do not guarantee that their systems will detect pedestrians because their sensing techniques may not return a strong reflection from people. Video- based recognition can be used to detect pedestrians. CWS, on the other hand, warn drivers of potential collisions by monitoring the time to contact with an object ( not including pedestrians). Forward CWS warn the driver of an impending collision with another vehicle or hard object. Rear CWS are fixed to the rear of the bus and warn other drivers if they are approaching too fast of an impending collision with the bus ( the bus operator would not be warned by rear collision warning). Finally, lane departure warning systems ( LDWS), although not technically designed to detect impending collisions, use image recognition to warn drivers of impending un- signaled lane departures that can lead to collisions or road departures.
• While a variety of vendors currently supply in- vehicle safety systems to the auto and heavy- truck industries, only two vendors have shown interest in pursuing the transit market. Of these two, only one has deployed its products to a limited number of U. S. transit operators. The effectiveness of each of these manufacturers’ products in reducing transit bus collisions remains undetermined.
• The study found that potential suppliers are hesitant to make any significant investments in developing transit- oriented products given the small transit market size and uncertainty in the demand among potential customers. At least one supplier is waiting for the U. S. DOT to make a funding decision before committing its own resources to further development. Similarly, some suppliers are awaiting an expression of widespread, organized interest among transit agencies.
Benefit- Cost Analysis
From a purely financial standpoint, only one of the standalone devices, SODS, was found to be cost effective ( i. e., the benefit- cost ratio exceeded one) under most, but not all, circumstances. While pedestrian detection systems were also found to be cost effective for operators with above- average collision rates or high collision costs, none of the seven standalone technologies evaluated here were found to be cost- effective under all circumstances. When bundled together as “ IVBSS” investments,
Federal Transit Administration 2
Assessing the Business Case for Transit IVBSS Final Report
systems containing a SODS performed best, generally passing the benefit- cost test under most conditions. However, as with the standalone systems, none of the bundled systems was able to pass the benefit- cost test under a full range of sensitivity assumptions.
The relative cost- effectiveness of SODS is driven by the fact that a high proportion of sideswipe collisions with other vehicles and collisions with fixed objects are avoidable by transit operators. This collision type is also relatively common, even though its costs are relatively low. On the other hand, the frequency and avoidability of forward, rear, and angle collisions is low, which hurts the potential for savings from technologies that address those types of collisions. Even with sensitivity analysis, few of these technologies demonstrated benefit- cost ratios above one.
Table 1- 1 and Table 1- 2 summarize the benefit- cost ratios for each device and six combinations of devices. The tables also include a range of ratios using 90- percent confidence intervals for the input variables. Even when considering the range, only SODS ( highlighted in the table in yellow) and combinations containing SODS have a high- end estimate above one. While advanced safety systems are appropriate and financially justifiable for high- speed applications such as over- the- road trucks and passenger cars, they do not appear to be as beneficial in the transit operating environment, where low- speed, low- impact collisions are often unavoidable.
Table 1- 1: Range of Benefit- Cost Ratios for Standalone Safety Systems1
In- Vehicle Safety System Description
Baseline
Range
Estimated purchase price
Forward Collision Warning ( FCWS)
0.45
0.22 - 0.81
$ 1,500
Rear Collision Warning ( RCWS)
0.59
0.10 - 1.44
$ 1,449
Side Object Detection ( SODS)
1.43
0.37 - 3.55
$ 2,550
Forward Object Detection ( FODS)
0.26
0.13 - 0.45
$ 2,350
Rear Object Detection ( RODS)
0.14
0.05 - 0.28
$ 2,550
Lane Departure Warning ( LDWS)
0.10
0.04 - 0.20
$ 900
Pedestrian Detection ( PDS)
0.81
0.11 - 1.62
$ 1,800
Table 1- 2: Range of Benefit- Cost Ratios for Integrated ( IVBSS) Systems2
Package of systems
Baseline
Range
Estimated purchase price
SODS & FODS
1.60
0.47 - 3.79
$ 2,750
SODS, FODS, & RODS
1.59
0.48 - 3.72
$ 3,150
SODS & FCWS
1.25
0.37 - 2.93
$ 4,250
FCWS & LDWS
0.49
0.22 – 0.88
$ 1,800
FCWS, LDWS, & PDS
0.83
0.21 - 1.61
$ 3,600
FCWS, LDWS, PDS, & RODS
0.69
0.19 - 1.33
$ 5,450
All 7 Systems
1.38
0.37 - 3.08
$ 7,999
When all seven systems are taken together, it is estimated that they can prevent approximately 22 percent of all transit collisions. If all 72,000 transit buses in the United States were equipped, this level of collision reduction would save approximately 15- 20 lives, prevent approximately 1,500 injuries, and reduce collision- related costs by nearly $ 100 million each year. SODS alone would prevent an estimated 11 percent of all collisions, thereby saving about 2- 3 lives, preventing 400 injuries, and reducing costs by over $ 40 million on an annual basis.
1 Results of benefit- cost analysis in 4.0.
2 Results of benefit- cost analysis in 4.0. Federal Transit Administration 3
Assessing the Business Case for Transit IVBSS Final Report
The benefit- cost analysis rests on several assumptions, including the following two key assumptions:
• First, the analysis assumes that in- vehicle safety technologies will be most effective in addressing incidents that the participating transit agencies considered “ avoidable” on the part of the bus operator. Table 1- 3 shows that avoidable collisions represent less than one- third of all collisions. This fact has major implications for in- vehicle technologies, which are assumed to be relatively ineffective against the two- thirds of collisions considered “ unavoidable.” In practice, one or more technologies may prove effective in preventing or mitigating some collisions currently considered “ unavoidable.”
Table 1- 3: Percent of Collisions Judged “ Avoidable” by Type of Collision3
With vehicle
Front
Rear
Angle
Sideswipe
Other
With pedestrian
With object
All collisions
28%
16%
14%
18%
32%
35%
90%
29%
• Second, the actual effectiveness of each of these technologies in preventing or mitigating bus collisions has not yet been determined. Given the absence of extensive service histories for these technologies, the study relied on lengthy reviews of over 60 well- documented collision scenarios to assess how each technology is expected to perform under each of these scenarios. Once again, the service histories of each of these technologies will need to be extensive, covering thousands of service miles before prevention effectiveness is known with any accuracy. Until such time that the actual effectiveness is measured empirically, the cost- effectiveness of these systems can only be estimated using techniques such as that applied here.
Industry Outreach
There was significant interest in the prospect of reducing collisions, which transit agencies view as costly nuisances that can strain relationships between the agency and its customers, the public and lawmakers. However, many expressed skepticism about IVBSS for several reasons. First, few believed that any in- vehicle system could address collisions without providing a large number of false alarms, given the dense operating environment of transit buses. In addition, many transit buses are already equipped with a variety of technical enhancements, which discourages managers from endorsing the deployment of further add- on systems. Furthermore, transit operators expressed doubt or concern that there will be advances made in the small and fragmented transit market. To elicit a reasonably priced, effective product offering from suppliers, agencies must organize and exert a coordinated demand.
That said, agency staff universally recognize the importance of safety as part of their ability to deliver services to customers, and all were willing to consider adoption of systems that can provide for meaningful collision reduction.
1.4 Recommendations
In- vehicle safety systems have the potential to deliver significant non- financial benefits to operators— most notably, an improved public image. Based on the results of the financial analysis, only SODS ( or packages containing side object detection) were found to be cost- effective under common operating conditions. Furthermore, at current avoidable collision rates and system prices, none of the other safety systems were found to be cost effective unless assumptions about collision occurrence and prevention rates were modified. However, pedestrian detection devices were determined to be cost effective for operators with an above- average number of pedestrian collisions and/ or high pedestrian collision costs.
3 Based on data collected from 2 transit agencies for the benefit- cost analysis in 4.0. Federal Transit Administration 4
Assessing the Business Case for Transit IVBSS Final Report
Moreover, many of the combinations of devices were cost effective, which suggests that an economy of scales exists, at least for those devices using common underlying technology elements.
Based on these findings and the feedback received from transit operators, the following recommendations are proposed.
Pursue Operational Tests on Those Systems with Higher Cost Effectiveness
Based on the results of this study, it is recommended that the U. S. DOT pursue further operational tests of the systems that appear cost effective. Namely, systems that address object collisions, sideswipes, and pedestrian collisions have the greatest potential to achieve substantial benefits. Among the existing systems, future evaluations should focus on SODS and pedestrian detection systems as likely the most ( and potentially the only) cost- effective standalone investments. Due to the low marginal cost associated with expanding the detection capabilities of a SODS to include forward and rear detection, the benefits of such an expansion may be compelling.
Determine True Effectiveness of Systems through Operational Tests
At present, there are no accurate empirical measurements of the effectiveness of bus collision avoidance systems in preventing or mitigating bus collisions. In the absence of such information, the benefit- cost analysis in this study rests on estimated system effectiveness rates based a detailed classification of accident scenarios. It is recommended that the U. S. DOT conduct sufficient operational tests to determine the effectiveness of these systems. Given the results of this study, these tests should again focus primarily on SODS and secondly on pedestrian detection systems as likely the most cost- effective systems. Once the collision prevention effectiveness of these systems has been assessed with sufficient accuracy, the benefit- cost analysis presented here should be updated. The U. S. DOT may also wish to conduct more limited testing of the remaining collision avoidance systems ( although the lower frequency collision reduction rates and lower cost savings expected with these systems may make it difficult to obtain a definitive assessment of these systems’ overall effectiveness).
Develop a Comprehensive Operational Test and Deployment “ Roadmap”
Develop a comprehensive operational test and deployment “ roadmap” similar to that outlined in 6.0. This roadmap and its related standards for designing and implementing operational tests will help ensure that test cases are well considered ( e. g., using proper control group comparisons), results are properly measured, and the findings are robust.
Focus on Human Interface Components
Further development of any in- vehicle systems for transit should consider additional improvements to system human interface components to minimize operator interaction requirements, maintenance needs, and false alarms.
Integrate Existing Bus Systems
Focus transit technology resources on integrating existing bus systems to make the acquisition of systems more efficient, simplify the level of technical sophistication required by agency operations and maintenance staff, and reduce the number of operator distractions, allowing them to focus their attention on their core competency— operating a motor vehicle.
Deliver a Consistent Message to the Transit Industry
During on- site interviews and a roundtable session, managers at many agencies expressed concern that the U. S. DOT’s progress in helping to develop, test, deploy, and encourage safety systems in the transit
Federal Transit Administration 5
Assessing the Business Case for Transit IVBSS Final Report
market appeared slow, which has caused some agency decision- makers to question whether this is a sign that there is little value in studying and deploying safety devices. Should the U. S. DOT decide to invest additional resources in the development of IVBSS for transit, it is imperative to deliver a consistent message to agencies on the level of federal commitment to the program, and to communicate progress regularly to the industry so that agencies do not draw inaccurate conclusions about the efficacy of safety systems.
Federal Transit Administration 6
Assessing the Business Case for Transit IVBSS Final Report
2.0 INTRODUCTION
2.1 Background
Roadway collisions take a significant toll on society. In 2005, there were over 42,000 fatalities and 2.5 million injuries in over 6.2 million police- reported crashes. This problem extends to transit buses, which annually are involved in over 15,000 federally reported collisions, leading to approximately 100 fatalities and 7,500 injuries. Moreover, collision records from six transit agencies participating in this study suggest that approximately 85,000 additional minor, property- damage- only ( PDO) collisions go unreported each year.
Proportional to the number of passenger- miles traveled, the likelihood of injury or death in transit bus- involved crashes is far smaller than for any other mode of roadway travel. However, the rate of collision incidents is nearly four times higher for transit buses than for all modes, as shown in Table 2- 1. Moreover, the mix of collision types for transit buses is different from the collision mix experienced by other roadway modes, as shown in Table 2- 2. Notably, the proportion of transit bus collisions that involves sideswipes is far greater than that of all roadway travel modes.
Table 2- 1: Safety Summary for Transit Buses vs. All Roadway Modes ( PMT= Passenger- Miles Traveled)
Transit buses4
All roadway modes5
Number
Rate per million PMT
Number
Rate per million PMT
PMT ( millions)
22,000
n/ a
4,700,000
n/ a
Crashes
100,000
4.55
6,200,000
1.32
Fatalities
80
0.004
43,000
0.009
Injuries
7,500
0.34
2,700,000
0.57
Table 2- 2: Mix of Collisions by Type for Transit Buses vs. Other Modes6
Collisions with vehicles
Mode
Front/ rear
Angle
Sideswipe
Other/ non- collision7
Pedestrian collision
Object collision
Transit Bus
30%
11%
33%
13%
2%
10%
All Roadway Modes
30%
29%
8%
4%
2%
27%
For transit agencies, however, the safety problem goes deeper than collision, fatality, and injury rates. Agencies must also address the financial costs directly associated with collisions, including accident repairs, claims payments, legal fees, workers’ compensation, and lost productivity. Collisions can also disrupt bus service and cause delays for all roadway users, inhibiting the operator’s ability to fulfill its mission to the public. Furthermore, public perceptions of safety can be tarnished by a single incident, thus eroding the trust and confidence of riders, lawmakers, and funding agencies and generating unfavorable media attention.
A systematic strategy for improving bus safety at transit agencies must begin by building an understanding of the causal factors that contribute to vehicle collisions. Such factors include bus operator errors, errors by other drivers, weather, infrastructure conditions, and vehicle conditions. Strategies
4 Source: Estimated based on 2005 National Transit Database ( NTD) and collision data collected from a sample of six agencies for this report.
5 Source: Federal Highway Administration ( FHWA) 2005 Highway Statistics ( Table VM- 1) and National Highway Transportation Safety Administration ( NHTSA) Traffic Safety Facts ( 2005 Data Overview).
6 The figures in this table corresponding to the “ transit bus” mode are based on analysis of data from NTD and collision records provided by six transit agencies participating in this study. The values shown for “ all roadway modes” are based on General Estimates System ( GES) data.
7 Includes, for example, road departure crashes, rollovers, and other accidents not otherwise classified. Federal Transit Administration 7
Assessing the Business Case for Transit IVBSS Final Report
targeted at reducing collisions must recognize the causal factor that they can influence. Examples of countermeasure strategies for these factors include driver training, deployment of on- board safety devices to assist operators with detection and warning of collision threats, and proactive vehicle maintenance. Figure 2- 1 illustrates these causal factors and countermeasure strategies.
Figure 2- 1: Collision Countermeasure Strategies Infrastructure conditionsDriver errorOther factors ( e. g., other drivers) Reduce crashesVehicle maintenanceWeather conditionsContributing crash factorsSample countermeasure strategies• Guardrails• Lane striping• Removal of roadside fixed objects and barriers• Removal of objects in roadway• Driver training• On- board safety technologies( e. g., collision warning and object detection systems) • Proactive maintenance strategies• Advanced vehicle diagnostic technologies• Proper vehicle maintenance• Driver training• Radio and other on- board comm- unicationsdevices• Driver training• Infrastructure technology deployment• Other strategies
Although each collision can have many causal factors, agency- supplied data indicate that between 10 and 30 percent of all transit bus- involved collisions are primarily the fault of the bus operator, while most of the remaining collisions are primarily the fault of other roadway users. Weather, infrastructure, and vehicle conditions also contribute to bus collisions, but rarely are the sole cause. Given these figures, operator error stands out as the largest single cause of bus collisions that the transit industry can address directly. Most transit agencies recognize this problem and have rigorous training processes for operators. These processes are designed to prevent accidents from occurring through operator training, while enforcing strict disciplinary and retraining procedures for operators who are involved in reported collisions.
At the same time, vendors of advanced technologies and university researchers have begun to develop, test, and market safety systems designed to assist transit bus operators in avoiding collisions with other vehicles, pedestrians, and fixed objects. These systems are in many cases also capable of helping to avoid collisions whose primary cause is not operator error. The purpose of this report is to present findings related to these systems, including their expected performance capabilities, cost effectiveness, and investment risks.
2.1.1 Federal Safety Technology Initiatives
In recognition of the need for improved roadway safety, including transit bus safety, the U. S. Department of Transportation’s ( U. S. DOT’s) Intelligent Vehicle Initiative ( IVI) began in the 1990s. The goal of the program was to improve vehicle safety for all modes of roadway travel through development and deployment of advanced in- vehicle safety systems. Since then, a variety of systems has been studied for application to passenger cars, heavy trucks, and transit buses. Specific systems evaluated under the IVI program include the following:
Federal Transit Administration 8
Assessing the Business Case for Transit IVBSS Final Report
• Forward collision warning
• Side object detection
• Rear impact collision warning
• Vehicle lane- assist
• Driver- vehicle interfaces ( DVI) for all safety devices
More recently, as the number and variety of on- board safety systems have expanded, vehicle manufacturers and operators have expressed interest in integrating them, in hopes of reducing investment costs while increasing the benefits of deployment. Recognizing the potential to improve the performance and accelerate the adoption of safety systems, the U. S. DOT initiated the IVBSS program. The U. S. DOT’s IVBSS program has the following goals:
• Integrate/ simplify technologies
• Increase safety benefits
• Improve overall system performance
• Reduce system cost ( e. g., due to economies of scale in system design or production)
• Improve acceptance of in- vehicle technologies among bus operators and transit management
• Enhance marketability of safety devices for transit.
Among passenger vehicles and heavy trucks, IVBSS has progressed to widespread demonstration and evaluation. Despite the progress in the auto and heavy- truck markets, adoption of in- vehicle safety systems by transit bus operators has been slow.
2.1.2 IVBSS for Transit
Two primary reasons likely account for the slow adoption of in- vehicle safety systems in transit. First, the size of the transit bus market is relatively small compared to the passenger car and heavy truck markets. There are approximately 72,000 transit buses in the United States, with average replacement cycles of 12 years or longer. By contrast, approximately 15 to 20 million new passenger cars are purchased each year, and the total fleet size now exceeds 230 million in the United States alone. The population of heavy trucks is well over 1 million, with average replacement cycles of 3 to 5 years. Given transit’s significantly smaller market size, many safety system manufacturers and vendors are understandably less interested in developing systems that meet the specific safety needs of transit operators.
Second, the operating environment of a transit bus is different from the driving environments generally experienced by other vehicle types. Specifically, transit buses tend to operate in densely trafficked urban settings, make frequent stops, interact often with passengers and pedestrians ( e. g., at bus stops and crosswalks), and operate at relatively low speeds ( i. e., average bus operating speeds are 12 mph). In contrast, autos and heavy trucks tend to travel at significantly higher average speeds, over many highway miles of travel, in lower traffic densities, and with less pedestrian interaction. Given that the vehicle safety systems evaluated under IVI were developed primarily with the larger auto and truck markets in mind, many are not well suited to addressing the types of threats typically encountered in the transit operating environment. For example, lane- departure warning, adaptive cruise control, and rollover stability control systems focus on preventing or mitigating crashes that occur most frequently in high- speed, long- haul environments.
Despite these challenges, a small number of suppliers has developed and is currently marketing safety systems whose functionalities address the unique requirements of the transit operating environment. University researchers have also developed prototype systems for testing. Many of these commercially Federal Transit Administration 9
Assessing the Business Case for Transit IVBSS Final Report
available products and research prototypes either have been tested or are currently being tested on transit agency vehicles. These commercially available products and research prototypes include the following safety system types, collectively referred to as Collision Avoidance Systems ( CAS):
• Forward Collision Warning System ( FCWS)
• Rear Collision Warning System ( RCWS)
• Side Object Detection System ( SODS)
• Forward Object Detection System ( FODS)
• Rear Object Detection System ( RODS)
• Lane Departure Warning System ( LDWS)
• Pedestrian Detection System ( PDS)
For purposes of this study, bus in- vehicle CAS have been separated into object detection systems ( ODS) and collision warning systems ( CWS). Figure 2- 2 illustrates this typology. As a rule, ODS monitor the area in close proximity to the vehicle and detect objects that are not within the view of the driver. Systems available at this time include front, side, and rear object detection. Most manufacturers of non- video- based ODS do not guarantee that their systems will detect pedestrians. Their sensing techniques cannot ensure return of a strong reflection from identified pedestrians. However, video- based recognition can be used to detect pedestrians. Therefore, for this study, pedestrian detection systems are considered a distinct category of ODS.
CWS, on the other hand, warn a driver of a potential collision by monitoring the time to contact with an object ( not including pedestrians). Forward CWS warn the driver of the equipped vehicle of an impending collision with another vehicle or hard object. Rear CWS are fixed to the rear of the bus and warn other drivers approaching too fast of an impending collision with the bus ( the bus operator would not be warned by rear collision warning). Finally, lane departure warning systems ( LDWS), although not technically designed to detect impending collisions, use image recognition to warn drivers of impending lane departures. When any two or more of these collision avoidance systems, whether CWS or ODS, are bundled together and “ integrated,” only then can they truly be considered an IVBSS investment.
Figure 2- 2: Typology of In- Vehicle Safety Systems CASCollision AvoidanceSystemsCWSCollision Warning SystemsForward ( FCWS) Rear ( RCWS) Lane Departure ( LDWS) ODSObject Detection SystemsSide ( SODS) Forward ( FODS) Rear ( RODS) Pedestrian ( PDS)
2.2 Study Objectives
The primary purpose of this study is to evaluate the business case for ( or against) the development of safety systems by equipment vendors and the adoption of those systems by U. S. transit bus operators. Specifically, the study addresses the question of whether the expected benefits from investing in these systems ( e. g., expected vendor sales or agency reductions in accident costs) outweigh the costs. Based on
Federal Transit Administration 10
Assessing the Business Case for Transit IVBSS Final Report
this study and other related analyses, the U. S. DOT will make a determination of whether further investment in transit- based versions of these systems warrant investment in additional operational tests, demonstrations, and evaluations. This report represents an input to that “ go/ no go” decision. Figure 2- 3 presents a schematic of the U. S. DOT’s decision- making process for further investment in transit IVBSS.
Figure 2- 3: U. S. DOT Decision- Making Process for Transit IVBSS Operational Testwith Vendor and SupplierWide Scale Demoand Evaluation20072008Assessment of transit marketDevelop deployment/ commercialization planValidation of specs and requirementsScenario developmentWorkshops/ discussions with stakeholders2005- 2006Go/ No GoGoNo GoStopToday
In addition to evaluating the financial benefits and costs of investment in safety systems, the study also addresses several questions of interest to potential IVBSS investors. These questions include:
1. What is the current state of the art? What types of safety systems are available, how do they work, and how effective are they likely to be in reducing collision frequency and collision severity?
2. Under what circumstances will transit operators deploy these systems? What characteristics do transit operators want these systems to have in terms of cost, maintainability, and human factors? How effective do the systems need to be before agencies will invest?
3. Are there any “ show stoppers”? Do transit operators harbor any key concerns or investment risks ( e. g., potential liability issues) that will effectively dampen or prevent widespread deployment of IVBSS and related systems?
2.3 Study Approach
This study adopted a three- pronged approach to evaluating safety systems as they apply to transit buses:
1. Technology Evaluation: Identified and evaluated the functional and technical characteristics of commercially available collision avoidance systems applicable to transit buses.
2. Benefit- Cost Analysis: Conducted a benefit- cost analysis of various safety systems ( both individually and in integrated “ IVBSS” packages) from the perspective of an investing transit agency to determine whether investment benefits exceed direct agency costs.
3. Industry Outreach: Conducted outreach sessions with transit operators to document their perceptions of these systems as well as the qualitative risks, rewards, and concerns that determine their interest in investing in these and similar systems.
Federal Transit Administration 11
Assessing the Business Case for Transit IVBSS Final Report
2.3.1 Technology Evaluation
This study’s technology evaluation provides a comprehensive review and assessment of all commercially available CAS and a brief assessment of potential future product offerings. The evaluation was completed using data from a wide range of sources including reports from prior technology evaluations and field tests, product supplier technology specifications, interviews with supplier representatives, interviews with users, and other research relating to in- vehicle safety systems. In addition to yielding a solid understanding of the differing objectives and operating principles behind each of these systems, this review provided an understanding of the expected effectiveness of these systems, which was essential to development of the study’s benefit- cost analysis.
2.3.2 Benefit- Cost Analysis
Each of the seven CAS is designed to reduce the frequency or mitigate the severity of one or several specific types of collision. For example, SODS are designed to reduce the incidence of sideswipe vehicle collisions and collisions with fixed objects. The cost- effectiveness of any given bus CAS is largely determined by the following four factors:
1. System Cost: What does it cost to purchase, operate, and maintain the system?
2. Collision Frequency: What types of collisions is the system intended to address and what is the frequency of those collision types? Also, what is the range ( and frequency) of severities for collision types addressed by the system?
3. Collision Cost: What is the typical cost of collision types addressed by the system ( for each level of severity)?
4. System Effectiveness: How effective is the system in reducing the frequency and/ or the cost of these collision types?
Obtaining answers to these questions is critical to the process of evaluating the business case for all seven bus CAS, both as standalone systems and as integrated investments. Systems will tend to perform best if they are both low cost and effective in preventing or mitigating accidents of high frequency and/ or high cost. For example, while the cost of collisions in which no one is injured is generally low ( relative to other collision types), the frequency of these collisions is very high. Consequently, any system that is reasonably effective in preventing or mitigating collisions of these types should also perform well from a financial perspective. In contrast, while fatal collisions are financially very costly, they are also extremely rare in the transit industry ( with fewer than 100 people killed in transit bus- related collisions each year, approximately one- quarter of whom are pedestrians). Because of the relative rarity of pedestrian injuries and fatalities, systems designed to reduce pedestrian collisions need to be very effective in preventing or mitigating such incidents in order to be financially effective.
It is also important to recognize that the interplay of the four factors identified above can vary significantly from one transit agency to the next. Each transit agency’s safety performance is unique and, by extension, the frequencies and costs of each collision type vary between agencies. Given this interplay, systems that are not cost- effective for some transit operators may prove highly effective for others. Similarly, system effectiveness could vary within a given agency if the operating environment varies sufficiently across that operator’s service area ( e. g., a system may be effective in dense urban traffic but not on suburban routes). The sensitivity analysis presented in below addressed these issues.
While recognizing the above issues, the objective of the study’s benefit cost- analysis was to determine which bus CAS— either individually or in integrated systems of two or more systems— appear cost- effective from the viewpoint of a transit agency ( i. e., the investment benefits exceed the investment
Federal Transit Administration 12
Assessing the Business Case for Transit IVBSS Final Report
costs). The benefit- cost analysis first assessed the frequency and cost of actual transit bus collisions by type and by severity. To do so, collisions were segmented into a matrix of seven collision types and five collision severities ( yielding 35 unique collision type/ severity combinations). Table 2- 3 presents a summary of collision rates.
Table 2- 3: Annual Frequency of Collisions by Type and Severity per 1000 Transit Buses8
Collisions with vehicles
Severity
Front
Rear
Angle
Side- swipe
Other
Collisions with pedestrians
Collisions with objects
Total
% of Total
Major fatal
0.2
0.2
0.1
0.0
0.0
0.4
0.0
0.9
0%
Major non- fatal
6
5
4
2
1
0
1
18
1%
Non- major with injury
15
12
8
3
2
4
10
48
3%
Non- major PDO
36
28
20
8
5
4
10
111
7%
Not FTA- reported
161
198
138
494
193
21
143
1348
88%
Total
218
244
170
507
202
27
158
1526
100%
% of Total
14%
16%
11%
33%
13%
2%
10%
100%
This segmentation permits a more precise categorization and understanding of the costs associated with collisions of each type and severity than has been considered in prior benefit- cost analyses of these systems. Specifically, using this segmentation, each safety technology could be assessed based on its ability to prevent or mitigate the costs associated with those specific types of collisions that technology was intended to address ( e. g., side- object detection systems were assessed in terms of their ability to prevent or mitigate sideswipe collisions and collisions with fixed objects impacting the side of the bus). Data to support these estimates were obtained from Federal Transit Administration’s ( FTA’s) National Transit Database ( NTD) and from six transit agencies. The agencies were: 1) San Francisco Municipal Railway ( Muni), 2) Alameda County Transit Authority ( AC Transit), 3) Los Angeles County Metropolitan Transportation Authority ( Metro), 4) North County Transit District ( NCTD), 5) Chicago Transit Authority ( CTA), and 6) Greater Cleveland Regional Transit Authority ( GCRTA). The complete analysis assesses how the cost of investing in each safety system ( either individually or in integrated systems with more than one system) compares with the benefits.
2.3.3 Industry Outreach
Finally, the study included outreach sessions designed to obtain input from U. S. transit agencies on their perceptions, interest in, and concerns with safety systems. These sessions included both a multi- agency roundtable session conducted during the October 2006 American Public Transportation Association ( APTA) Annual Meeting in San Jose, CA, and extensive on- site meetings with management and staff of six different transit agencies from across the country. These interactions generated a wealth of
8 The number of collisions “ Not FTA- reported” was estimated based on data provided by agencies participating in this study. All other data were obtained directly from NTD. Federal Transit Administration 13
Assessing the Business Case for Transit IVBSS Final Report
information about agency perspectives on anticipated system deployment issues, including those relating to investment risks, driver acceptance, anticipated system effectiveness, experiences with new system investments in general, and their interest in IVBSS in particular.
2.4 Report Format
The remainder of this report provides detailed results of the research and analyses. Chapter 3 “ Technology Assessment” summarizes safety systems for transit buses, including those currently available and those under development. Chapter 4 summarizes the benefit- cost analysis and results. Chapter 5 discusses the “ soft business case” based on qualitative feedback from a sample of transit agencies that participated in this study. Chapter 6 suggests steps for future study and deployment. Finally, Chapter 7 summarizes the findings and recommendations for transit IVBSS.
Federal Transit Administration 14
Assessing the Business Case for Transit IVBSS Final Report
3.0 TECHNOLOGY ASSESSMENT
The objective of this chapter is to provide a comprehensive assessment of the types of collision avoidance systems ( CAS) that are either commercially available or for which prototype systems exist. Table 3- 1 provides an overview of the reviewed technology. The table depicts the categories of systems used for analysis in this study, the sensor technologies used in each system, the companies currently manufacturing each system, and a description of the commercial availability of each system.
The sections in this chapter correspond with the column headings in Table 3- 1. The first section begins with a description of each CAS. The next section reviews the general types of sensor technologies underlying the various existing CAS. Section 3.3 then provides detailed descriptions of the two commercially available CAS suitable for use in transit bus operations. Section 3.4 describes the three prototype CAS used in transit operational tests and provides an overview of safety systems available in the heavy- truck market. Finally, Section 3.5 discusses collision avoidance systems in the truck market ( not represented in Table 3- 1).
Table 3- 1: Overview of Collision Avoidance Systems
Collision Avoidance Systems
Sensor Technology
System Manufacturer
System Availability
Video
Mobileye
Commercialized
FCWS
Lidar/ Radar
PATH
Prototype
RCWS
Lidar
AATA / UMTRI
Prototype
Ultrasonic
Clever Devices
Commercialized
SODS
Video
Mobileye
Under development
FODS
Ultrasonic
Clever Devices
Commercialized
Ultrasonic
Clever Devices
Under Development
RODS
Video
Mobileye
Under development
Video
Mobileye
Commercialized
Video
Iteris
Commercialized - trucks only
LDWS
Video
Assistware
Commercialized - trucks only
PDS
Video
Mobileye
Application only
The key findings of this chapter include the following:
• Only two CAS are currently commercially available in the transit industry.
• Only the RCWS is envisioned as a standalone product; the remaining CAS either exist as part of or are planned to be part of a bundled package of two or more integrated systems.
• The dominant technologies used for CAS are video and ultrasonic detection. Video detection uses cameras together with object recognition algorithms and software to identify potential threats to vehicle safety, while ultrasonic detection uses radar waves to identify the location and proximity of objects.
Federal Transit Administration 15
Assessing the Business Case for Transit IVBSS Final Report
3.1 Collision Avoidance Systems
In 1990, a little- known research and development company, Radar Control Systems, debuted one of the first forward collision warning applications in anticipation of a new and emerging market for intelligent vehicles. The system used a radar transmitter/ receiver to scan traffic in front of the vehicle. By processing collision warning algorithms, the application was capable of predicting an impending crash and aiding the vehicle in reacting prior to the collision. Since 1990, CAS expanded to include a broad array of additional applications, including FCW, LDW and automatic cruise control, all designed to assist drivers and improve safety. CAS provides the drivers with knowledge of the environment surrounding the vehicle with the intention of reducing the probability of accidents. Figure 3- 1 depicts the “ zone” of applicability of the various CAS. CAS can be divided into two basic categories: 1) object detection systems and 2) collision warning systems. Each system is described in detail below.
Figure 3- 1: Diagram of Collision Avoidance Systems ( Plan View of a Bus Facing to the Right)
FCWS / PDSRCWS FODS SODS RODS LDWS PDS RCWS FCWS / PDS RCWS RODS L D W SSODSFODSLDWS S O D SL D W SLDWS
3.1.1 Object Detection Systems
Object detection systems ( ODS) are intended to monitor the area within close proximity of the vehicle ( e. g., up to 10 feet) and provide a visual or audible warning when an object is detected near the vehicle. Given this small proximity, ODS are sometimes considered an “ enhancement” to the driver’s mirror. These systems can detect the presence of an object but not its distance or relative speed. In Figure 3- 1, ODS are represented by the circular shaped areas. They are defined as follows:
• FODS – As shown by the gray/ light blue- shaded area, FODS monitors the area in front of a vertical plane intersecting the front bus wheels ( the area within the forward view of the driver).
• SODS – As shown by the dark green area, SODS monitors the area behind the vertical plane intersecting the front bus wheels. It does not include the area behind the vehicle, only the area from the front wheels, down the side of the bus, to the rear bumper.
• RODS – As shown by the dark blue area, RODS monitors the area directly behind the vehicle.
Federal Transit Administration 16
Assessing the Business Case for Transit IVBSS Final Report
3.1.2 Collision Warning Systems
Collision warning systems ( CWS) monitor distances further away from the vehicle ( up to 500 feet) and warn the driver of impending collisions. Algorithms use distance and relative speed information supplied by the detection sensors to calculate the time to contact a detected object, and then provide the driver with visual or audible warnings that increase in intensity as the time to contact approaches zero. The systems will provide warnings to the driver as vehicles/ objects enter the field of view or as the vehicle approaches a fixed object.
Figure 3- 1 illustrates the different type of CWS with triangular shapes. They include the following:
• Forward Collision Warning System ( FCWS) – Shown in yellow, FCWS uses forward- looking sensors and warns the driver of the “ Time to Contact” with a vehicle in the driver’s lane. Forward sensors are situated in the front of a vehicle with a widening view as they scan farther ahead.
• Rear Collision Warning System ( RCWS) – Shown in red, RCWS warn the driver of an approaching vehicle of a rear- end collision. The warning is an external indicator on the back of the equipped bus that alerts the driver in the approaching vehicle. ( The driver of the equipped vehicle is not alerted of the impending collision.)
• Lane Departure Warning Systems ( LDWS) – As shown by the green area, LDWS are camera- based systems that monitor lane markings. Together with object recognition software and algorithms that compute closing distance, LDWS provide warnings when a lane or road edge departure is imminent via visual, audible, or tactile warning signals.
3.1.3 Pedestrian Detection System
Pedestrian detection systems ( PDS) notify the vehicle operator of an impending collision with a pedestrian. The systems can be designed to provide cocoon or direction of travel coverage. Due to constraints with radar and Lidar sensors, video- based recognition accounts for the majority of technologies used to implement pedestrian detection. The systems use pattern recognition and optical flow techniques to differentiate between a pedestrian and an inanimate object. PDS detect pedestrians through a search of objects containing specific characteristics. The systems then separate a potential pedestrian from the background images. The software compares body ratios, specific size constraints, etc. to differentiate a non- human object from a pedestrian. A PDS has a normal range of 10 to 40 meters.
3.2 Sensor Technologies
Each collision avoidance system relies on at least one of the following four underlying detection technologies:
• Lidar, which are radar- like systems that function at near- infrared wavelengths
• Traditional radar- based systems
• Ultrasonic- based sensors
• Video- based systems
The role of these technologies is to provide information on the presence of objects near a vehicle, the proximity of those objects and, for some technologies, the differences in the relative speeds of the bus and the detected object. The selection of which specific detection technology to use in developing any given collision avoidance system depends directly on the system’s intended application, the desired performance characteristics, and the supplier’s design philosophy. The following are brief descriptions of each of these detection methods.
Federal Transit Administration 17
Assessing the Business Case for Transit IVBSS Final Report
3.2.1 Lidar- Based Systems
Lidar- based systems transmit a light beam to the area surrounding the vehicle and then detect the presence of nearby objects through the reflected signal. In addition to direction, Lidar systems can determine an object’s distance and relative speed. The ideal operation range for Lidar is 2 to 30 meters over which this technology provides excellent angle resolution. Lidar systems are susceptible to the weather conditions ( e. g., to being able to “ see” through fog or heavy precipitation). In general, if an object is not “ detectable” by the naked eye, it is unlikely that a Lidar- based system will provide an adequate warning of an impending collision. Therefore, during times of fog, heavy rain, or heavy snow, the system will become inoperable. Given these characteristics, Lidar- based systems are preferred by those that believe a collision avoidance system should not extend beyond the driver’s view. This position is based in part on the concern that systems that extend the driver’s view beyond what is visible with the naked eye may encourage reckless driving, particularly in poor weather conditions. Lidar sensors have a high cost of implementation and the output power level must be limited to meet eye safety constraints due to the light beam operating in the near- infrared range.
3.2.2 Radar- Based Systems
In contrast to Lidar, the performance of radar- based systems is not adversely affected by poor weather conditions. Hence, this technology is favored by those who believe collision avoidance systems offer their greatest benefits during adverse weather. Radar- based systems are capable of detecting objects out to 150 meters but suffer from low angular resolution, poor detection at medium range ( i. e., 30 to 60 meters), and generally inferior resolution to Lidar. As with Lidar, radar sensors have a high cost of implementation.
3.2.3 Ultrasonic- Based Sensors
Ultrasonic- based sensors are reliable and inexpensive. They operate at a high frequency ( 20 kHz to 200 kHz) and are similar to the back- up sensors installed on sports utility vehicles. The sensors emit an ultrasonic signal that is capable of traveling 10 to 12 feet. The system detects the object when a recognizable echo is reflected from it and can measure the detected object’s distance and relative speed. Sensors provide a clear signal for detection algorithms and are less influenced by interference than are radar and Lidar. Their disadvantage is the limited detection range; they cannot detect objects beyond a small area around the vehicle. In addition, they are only capable of providing a recognizable echo from solid objects. Therefore, they should not be used for “ soft object” detection ( e. g., pedestrians).
3.2.4 Video- Based Sensors
Video- based sensors use a forward- looking camera for detection of objects. A pixel- based recognition algorithm identifies objects that may be of concern to the driver. The use of pixel- based recognition can distinguish pedestrians from other objects, a form of detection that is not possible with Lidar, radar, or ultrasonic- based systems. With the low- cost of the camera, video- based sensors have a low cost of implementation. Video- based sensors rely on ideal lighting conditions for detection. Therefore, in situations where the driver’s field of vision is impaired, the system will not function well ( including adverse weather conditions, direct sunlight, evening). In the video systems reviewed below, the video- based suppliers supplemented their systems using infrared sensors to ensure object detection under a greater range of conditions than that permitted by a video- based system alone.
3.3 System Manufacturers
Market research conducted for this study only identified two companies interested in supplying collision avoidance systems to the transit bus market. Beyond these two suppliers, commercial and passenger
Federal Transit Administration 18
Assessing the Business Case for Transit IVBSS Final Report
vehicle suppliers have shown little interest in expanding their line of business to include transit buses given the small size and specialized needs of the transit market. A number of commercial companies were contacted to determine the reasons for lack of interest. This information is in the section entitled, “ Prototype Systems.”
3.3.1 Clever Devices
Clever Devices has focused solely on providing technology solutions to the transit industry since 1987. The company’s products provide improved communications and safety systems for the transit agency applications, including passenger information systems and intelligent vehicle systems. Clever Devices entered the ODS market in a partnership with the FTA, Carnegie Mellon University, and the Port Authority of Alleghany County. They developed a prototype as part of the FTA’s Intelligent Vehicle Initiative. The original product, the Enhanced Object Detection System, was refined during the IVI field tests and commercialized as the Seymor System.
Clever Devices’ Seymor System is marketed specifically for object detection within a transit bus application. The system was designed to be an extension of the driver’s mirrors, providing blind- spot coverage. Ultrasonic sensors detect non- stationary objects within a defined perimeter ( Figure 3- 2). The sensors are installed at six locations on the vehicle— one sensor on each front corner and one sensor each fore and aft of the left/ right front wheels ( Figure 3- 3). Sensors may also be installed at the rear of the vehicle for backing functions. The sensors transmit a signal and detect objects based on a recognizable echo reflected from an object.
Figure 3- 2: Clever Device’s Ultrasonic Sensor9
9 Source: Clever Devices – Installation, Operation, and Maintenance Instructions for Seymor Object Detection System. Revision 1.1, November 2004. Federal Transit Administration 19
Assessing the Business Case for Transit IVBSS Final Report
Figure 3- 3: Seymor System Sensor Locations10
The system employs three distinct modes of operation ( Figure 3- 4) based on the vehicle speed:
1. Mode 1 – When the system is operating in an urban/ slow environment ( 0 to 15 mph), the system will detect objects within a 4- foot perimeter of each sensor. All sensors within the system will be active. If the system detects an object, a visual aid will flash with a frequency based on object distance. ( In Figure 3- 4, the yellow area equates to 4 feet.) As the object moves closer to the vehicle, the frequency of the blinking light increases. Finally, an audible tone will sound when the object is within 2 feet of the vehicle.
2. Mode 2 – When the system is operating in an urban/ fast environment ( 15 to 45 mph), the system will detect objects only when a turn signal is activated. With an activated turn signal, the detection zone is a 6- foot perimeter of the side sensors in the direction indicated by the turn signal. The front sensors are inactive at speeds over 15mph. If the system detects an object, it issues a solid visual indicator in conjunction with an audible alarm. ( Note: Figure 3- 4 shown with right turn signal activated.)
3. Mode 3 – When the system is operating in a highway environment ( 45+ mph), the system will operate similar to the Mode 2 with the exception of a detection zone of 8 feet for the activated side of the vehicle. ( Note: Figure 3- 4 is shown with right turn signal activated.)
Figure 3- 4: Seymor Modes of Operation11
The Seymor System communicates object detection through visual and audible warnings to the driver ( Figure 3- 5). Three identical visual driver displays are mounted within the peripheral line of sight of the side mirrors ( one at the left mirror, and a high/ low mounting at the right mirror; see Figure 3- 6). As the
10 Ibid.
11 Ibid.
Federal Transit Administration 20
Assessing the Business Case for Transit IVBSS Final Report
operator uses the mirrors for turning maneuvers or lane changes, he/ she can reference the displays to determine whether unseen objects are present. A flashing display ( frequency determines distance of object to vehicle) will notify the operator of a potential object in the path of the vehicle. A speaker is mounted behind the driver’s seat to provide audible warnings as the threat of the object increases.
Figure 3- 5: Seymor Driver Visual Interface ( DVI) 12
Figure 3- 6: Seymor DVI Locations13
The Seymor System has the following properties and characteristics:
• Detection Options: The Seymor System is designed for front and side object detection as the standard model. Clever Devices is in the development stages for a RODS. Combined with the standard system, the RODS would provide a cocoon surrounding the vehicle for detecting objects. A standard system ( front only) costs $ 2,600 to $ 2,900 ( excluding engineering design customizations). The system includes a standard 1- year warranty and maintenance option. The projected lifespan of the technology is 10 to 15 years. The RODS is not available for commercial sale at this time.
• Applicable Properties: The first generation Seymor System has been installed at the Washington Metropolitan Area Transit Authority ( WMATA), the Utah Transit Authority ( UTA), the Greater Cleveland Regional Transit Authority ( GCRTA), and the Port Authority of Allegheny County ( PAAC). Delivery time for the system is approximately 8 to 10 weeks. Installation requires 4 hours per bus. Clever Devices has established a training course for the Transit Authorities for operators and maintenance personnel. The classes take from 0.5 to 1 day. As the system is very intuitive, the training required is minimal.
• Business Plan: Clever Devices markets the Seymor System with its corporate capabilities. By attending trade shows and distributing brochures on the technology, Clever Devices has been
12 Ibid.
13 Ibid.
Federal Transit Administration 21
Assessing the Business Case for Transit IVBSS Final Report
marketing the advantages of the Seymor System for the transit agency. As of January 2007, WMATA has installed the system on 50 buses for a year- long trial. Other transit authorities have worked with Clever Devices to develop a system specification to require the installation of the Seymor System in their next vehicle purchase. Clever Devices has not been contacted by bus manufacturers to date, leading the company to believe the specification of their product was not part of the final purchase agreement by those other authorities.
• Future Technologies: Clever Devices is currently in the final stages of developing an RODS application. The system has entered the final testing phases and may be released to the market shortly. The application is an upgrade to the existing Seymor Forward Object Detection System. Four sensors would be installed at the rear of the bus to support back- up object detection notification. Cost data for the application has not been released at this time.
3.3.2 Mobileye
Founded in 1999, Mobileye’s mission has been to develop vision systems for accident reduction and driver assistance. They have established themselves as a leader in vision systems for intelligent transportation systems. Through the years, the company has developed algorithms and hardware for lane departure warning, headway monitoring, and collision mitigation applications. The technology has been installed as an aftermarket product under the AWS ( Advance Warning System) brand name. AWS integrates a series of advanced safety systems for installation as a single collision warning system. Mobileye’s products are used in the auto, truck, and transit markets.
Mobileye uses monocular vision analysis techniques to detect vehicles and to measure the distance and relative speed between vehicles and between vehicles and objects. The technique also measures the vehicle position relative to the lane boundaries as well as road geometry and lane curvature to identify the “ closest in path” vehicle. Mobileye uses a single video camera mounted on the front windshield and Mobileye’s EyeQ CMOS chip to detect objects.
Mobileye’s EyeQ product provides a low- cost solution while combining high performance and consolidating multiple applications on a single platform. Installed on a single board half the size of a standard business card ( Figure 3- 7), it processes visual images along three main areas— pattern recognition ( vehicles, pedestrians), image processing ( lane following), and visual motion understanding ( analysis of collision and cut- in maneuvers). The classification of these visual images allows the system to assist in preventing unintentional lane departure, detecting forward collision scenarios and maintaining a safe headway.
Figure 3- 7: EyeQ Chip and SeeQ Board14
Mobileye’s AWS is an aftermarket driver assistance system for accident prevention and mitigation. It combines the benefits of forward collision warning, lane departure warning, and headway assistance in a
14 Reprinted with permission from Mobileye Vision Technologies, Ltd. Federal Transit Administration 22
Assessing the Business Case for Transit IVBSS Final Report
single package. A single camera is mounted on the front windshield along with the EyeQ processing board ( Figure 3- 8). The package detects and measures distances to lanes and vehicles to provide timely alerts of impending safety- critical situations.
Figure 3- 8: Mobileye's Monocular System - Single Camera with Internal EyeQ Board15
The Forward Collision Warning ( FCW) application detects situations where the vehicle has the potential to collide with another vehicle if no change is made to the speed or direction of travel. Using EyeQ’s vision- based algorithms, the system determines an object’s boundaries and classifies the target as vehicles or non- vehicles. If the object is determined a threat, the system tracks the time to contact ( TTC) ( Figure 3- 9). As the TTC falls below 2.7 seconds, the system begins to issue a series of warnings to the driver. The system will continue to monitor the TTC, continuing the alert if the driver does not react to the initial warnings. The application operates at speeds above 3 mph.
Figure 3- 9: Mobileye's DVI Demonstrating FCW16
The Lane Departure Warning ( LDW) application uses a lane detection algorithm to detect lane markings and provide various measurements related to them. The system is capable of detecting a variety of lane markings ( e. g., solid markings, dashed markings, and double lane markings) under various weather and road conditions ( e. g., asphalt, concrete). The color of the markings and the time of day do not affect the technology. The algorithms measure the distance from the vehicle’s wheel to the marking. The vehicle speed with respect to the lane marking is calculated from the vehicle's lateral position, lateral speed, road curvature, and speed. The system will warn the driver of an impending lane departure only if the appropriate turn signal has not been activated. The warning associated with LDW is both an audible ( direction rumble sound relative to direction of deviation) and visual ( Figure 3- 10). The figure demonstrates a left- side lane departure. The system setting is adjustable so that the system will issue a warning when the vehicle crosses the lane marker or when it approaches the lane marker. The application is only active at speeds over 34 mph.
15 Ibid.
16 Ibid. Federal Transit Administration 23
Assessing the Business Case for Transit IVBSS Final Report
Figure 3- 10: Mobileye's DVI Demonstrating LDW17
The Headway Monitoring and Warning ( HMW) application provides the driver with a digital distance gauge to assist in keeping a safe separation from the vehicle in front. The headway algorithms identify the rear profile of a car in lit situations and the rear taillights in the evenings/ unlit conditions. The information detects the closest vehicle in the path of the driver. The vehicle display will provide a headway distance to the detected vehicle, in seconds, after the separation has fallen below 2.5 seconds. As the vehicle closes in on the detected vehicle, the seconds on the display will decrease and the icon will change from green to amber to red ( Figure 3- 11). Once the headway has reached a dangerous separation, an audible warning will alert the driver. The timing of the audible warning is adjustable based on the level of security the driver desires ( early versus late warning). The HMW application is only active at speeds over 25 mph.
Figure 3- 11: Mobileye's DVI Demonstrating HMW18
Mobileye offers a separate Pedestrian Detection technology/ algorithm. Mobileye currently provides pedestrian detection as a technology only. The customer is responsible for implementing all applications, including driver interface and detection hardware. The technology uses the monocular vision camera and infrared sensors for night applications ( Figure 3- 12).
Figure 3- 12: Pedestrian Detection19
17 Ibid.
18 Ibid.
19 Ibid.
Federal Transit Administration 24
Assessing the Business Case for Transit IVBSS Final Report
The pedestrian detection application is one of several advanced development programs for production by automotive original equipment manufacturers ( OEMs). Using the basic monocular camera ( as used within the AWS- 4000 system), the algorithm can detect pedestrians based on the visual spectrum. To achieve daylight performance levels regardless of lighting conditions, a near infrared sensor is required. The system can effectively detect passengers/ pedestrians within a 30- meter range. In a future aftermarket application, the system will be paired with a visual/ audible alert for the driver. The visual alert is implemented through an LED or other icon on the driver’s display. An audible warning is annunciated through the existing speaker system. The mid- range detection option can be expanded to include 360- degree coverage. The all- around option is implemented with six individual cameras installed around the vehicle. The range of the detection is shortened to 15 meters, and is intended for scenarios with slow- moving vehicles. The driver is notified of a pedestrian as done with the mid- range detection.
The AWS- 4000 is a complete package for front collision warning, lane departure, and headway monitoring. The system is bundled and designed specifically for after- market installations. It includes the driver display, the camera/ processor combination, and speakers. The unit retails for $ 1,800 installed. The system has a 1- year warranty and maintenance option, with a lifespan of at least 5 years.
The pedestrian detection application is a sensing application only with the software running on the EyeQ chip or the SeeQ board. The integrators are responsible for providing the appropriate displays to warn the driver of a detected pedestrian. The estimated cost for implementation is $ 1,800. The system may be bundled with the AWS- 4000 for an additional cost.
In addition, Mobileye offers several individual applications to enhance AWS. These applications include:
• Night Vision ( Near/ Far Infrared) – The night vision is required for pedestrian detection. It also enhances the video analysis during the evening or during times of adverse weather.
• Side Object Warning – Using the forward collision warning components and software, cameras installed around the vehicle detect slow- moving vehicles. The detection distance is just 45 feet, but the field of view is 90 to 100 degrees. The system is under development at this time with an expected release date at the beginning of 2008.
• Blind Spot/ Lane Change Aid – Using cameras mounted in the side mirrors, the system analyzes the opportunity to change lanes. It will estimate speed of approaching vehicles and warn if the speed is excessive. The cameras are capable of detecting vehicles in the adjoining lane within 60 meters. The technology has an expected release date of the end of the third quarter in 2008.
The technologies offered by Mobileye could be combined to provide a complete object detection system for the area surrounding the vehicle.
The Mobileye system has been installed on 150 transit buses in Israel. At this time, Mobileye is establishing a distributor network. Installation time should not exceed four hours per vehicle. The system does not require driver interaction after installation, but allows driver customization ( e. g., volume, display brightness, and warning level). With a quick- start guide, it is estimated that the operator can begin using the AWS- 4000 within 10 minutes.
Mobileye is scheduled to provide its technology to six passenger car production platforms in 2007 ( start of production) with a major U. S. and two leading European car manufacturers, including the BMW 5- series. Additional OEMs and Tier 1 suppliers ( i. e., those supplying GM, Chrysler, and the other major auto manufacturers) have Mobileye evaluation systems installed in vehicles for advanced development and research programs with an additional number of production intent agreements in place.
Federal Transit Administration 25
Assessing the Business Case for Transit IVBSS Final Report
Within the next year, Mobileye plans the release of two new data logging technologies. By the end of the second quarter of 2007, a data logging feature will be available for the AWS- 4000. Owners will be able to track the driving practices of their operators, including typical headways, number of lane changes, number of illegal lane changes, etc. By the end of the third quarter of 2007, a video accident recorder will be available for the AWS- 4000. The video recorder is activated at the time of a collision. The video will record 20 seconds prior to the crash, and continue to record for 5 seconds following the crash. The video can be used for accident reconstruction.
Over the next two years, Mobileye plans to finalize two collision warning/ detection technologies. The Side Object Warning technology has an expected release date of the beginning of 2008. The Blind Spot Detection/ Lane Change Aid has an expected release date of September 2008.
3.4 Prototype Systems
Several prototype collision avoidance systems for transit buses have been developed and field tested in recent years. These include:
• Rear Impact Collision Warning System ( Ann Arbor Transportation Authority)
• Forward Collision Warning System ( California DOT)
• Side Collision Warning System ( Pennsylvania DOT)
• Integrated Collision Warning System ( California and Pennsylvania DOT)
• Pedestrian Warning Devices ( PDS)
The following sub- sections describe each of these systems and the known test results.
3.4.1 Rear Impact Collision Warning System
The Rear Impact Collision Warning System ( RICWS) 20 was developed by the Ann Arbor Transportation Authority ( AATA) in partnership with General Dynamics. The RICWS was a research project to assist the FTA in the mitigation of rear- end transit bus collisions. The system provides warnings to vehicles following a bus as the headway between the bus and the vehicle was reduced ( see Figure 3- 13).
Figure 3- 13: RICWS Installed at AATA21
The system used a rear- scanning Lidar sensor, a processing unit, and an 8- segement LED display. The Lidar sensor detects the presence and range of an approaching vehicle. The sensor was capable of detecting a vehicle up to 125 meters, with the vehicle detected before it is within 72 meters. A processing
20 Ann Arbor Transportation Authority and Kirk Lucksheiter. Develop Performance Specifications for a Rear Impact Collision Warning System for Transit Systems. November 2003.
21 Source: Final Report, Develop Performance Specifications for a Rear Impact Collision Warning System for Transit Systems. November 2003. Federal Transit Administration 26
Assessing the Business Case for Transit IVBSS Final Report
unit uses the information in conjunction with the bus’s speed to determine whether the warning criteria are met. If the processing algorithm determines evasive action may be necessary, the 8cm by 150cm LED is illuminated with increasing warnings. The system is capable of determining the approaching vehicle’s time to contact and provide sufficient warning to allow the approaching driver to brake/ swerve to avoid a collision. The RICWS is autonomous, and does not impose or distract the driver of the bus. The driver is not warned of the impending crash.
The system has not been commercialized at this time. The field operational test was completed in Ann Arbor on two transit buses in 2003. A specification has been developed for the system, but no further action has been taken.
3.4.2 Forward Collision Warning System
The University of California at Berkley’s PATH program developed the Forward ( Front) Collision Warning System ( FCWS) 22 under the direction of the FTA. PATH worked with the California Department of Transportation and SamTrans to design a system that could detect imminent crashes, provide warnings for smooth maneuvering, and provide warnings for reduced headways. The system, as outlined in the final report, contained five sensors, four cameras, and a single processing unit.
The FCWS contains radar and Lidar sensors to enhance the detection capabilities of the system. Two forward- looking radar sensors are on the right and left front corners of the bus ( see Figure 3- 14). A single forward- looking Lidar sensor is at the center of the bus. Two forward- looking Lidar sensors are installed in tandem with the radar sensors. The sensors measure the distance and angle to the detected object. The system is capable of detecting objects within the same lane from 3 to 100 meters. The recommended Lidar sensors are deactivated below 3 m/ s to ensure the safety of the surrounding pedestrians.
Figure 3- 14: FCWS Sensor/ Camera Locations23
22 X. Wang, et. al. Transit Bus Frontal Collision Warning System. March 2002. X. Wang, et. al. Development of Requirement Specifications for Transit Frontal Collision Warning System. August 2003.
23 Source: Final Report, Development of Requirement Specifications for Transit Frontal Collision Warning System. August 2003. Federal Transit Administration 27
Assessing the Business Case for Transit IVBSS Final Report
The cameras mounted as part of the FCWS were for recording capabilities only. The video feeds were not used for processing algorithms. Cameras were installed as follows: 1) one forward- looking camera, 2) one backward- looking driver and passenger- side camera, and 3) one internal ( passenger cabin) camera. The cameras in conjunction with the sensors identified detected objects during the field tests. The cameras may also be used as an accessory for reviewing injury claims.
The FCWS DVI is a set of vertical columns on the left and center pillars of the bus ( Figure 3- 15). The columns illuminate from top to bottom to indicate increasing severity. Each column independently operates to notify the operator of the physical location of the object. If the object is in front of the vehicle, both columns will illuminate simultaneously.
Figure 3- 15: FCWS DVI24
DVI
The system has not been commercialized at this time. A specification has been developed for the system, but no further action has been taken at this time. The design was provided to Mark IV for commercialization possibilities. At this time, they have not initiated plans to commercialize the system.
3.4.3 Side Collision Warning System
Carnegie Mellon University’s Robotics Institute developed the Side Collision Warning System ( SCWS) 25 under the direction of the FTA. The Robotics Institute worked with the Pennsylvania Department of Transportation and the PAAC to develop a system that tracked objects surrounding the bus ( within a 3- meter perimeter). The system is capable of detecting objects up to 50 meters away. It contains laser scanners for object detection and equipment for curb detection/ prediction. In addition to the sensors, the processor monitors the vehicle communications network to establish the bus states ( turn signals, speed, warning lights, door status, etc).
The scanners are installed on each side of the bus behind the front wheel and below floor level. They extend four inches from the side of the bus, but are installed within a retractable box. The scanner is capable of detecting an imminent collision, retracting prior to colliding with an object ( Figure 3- 16).
24 Ibid.
25 D. Duggins, et. al. Developing and Testing of Performance Specifications for a Next Generation Side Collision Warning System. April 2002. Federal Transit Administration 28
Assessing the Business Case for Transit IVBSS Final Report
Figure 3- 16: SCWS Retractable Laser Scanner26
The intention was for the scanners to predict and warn the driver of a pedestrian under the bus. When the bus was at a speed below 5 mph, the scanner tracks objects/ pedestrians as they approach the vehicle. If the pedestrian disappeared after entering a particular range surrounding the bus, the scanner would hypothesize that the pedestrian fell under the bus. To account for pedestrians boarding the bus, the algorithm used the door state of the bus to cancel any warnings that arose in the area of the door during embarking/ disembarking. During testing, the pedestrian detection algorithm was not successful. The laser scanner did not clearly differentiate pedestrians from inanimate objects. The original tests showed a number of false positives, suggesting that design enhancements are required prior to implementation.
The curb detector is part of the SCWS for curb prediction and detection. If the system detects a pedestrian on the curb, it is considered safer than if it detected in the roadway. It contains a laser line striper ( LLS) and a camera installed inside the front bumper on the non- driver’s side. The LLS projects a pattern of light that is imaged by the camera. The results are used to compute distance to detected objects. The system returns the cross- section profile of the environment beside the bus, providing the final distance from the bus to the curb.
The SCWS DVI illuminates a set of arrows to warn of a detected side object. The illuminated arrows represent the location of the object. For example, the top arrow illuminates when an object is detected at the front of the bus, and the bottom arrow illuminates when an object is detected at the back of the bus.
The current plans do not include the commercialization of the SCWS. Not only is cost for the system hardware is cost prohibitive, there are problems with the detection algorithms.
3.4.4 Integrated Collision Warning System ( Forward/ Side Collision Warning)
The Integrated Collision Warning System ( ICWS) 27 combined the research from PATH and the Robotics Institute to integrate the two separate collision warning systems into a single product. The teams continued their work with the California Department of Transportation ( Caltrans) and the Pennsylvania Department of Transportation. A complete system was installed on vehicles at SamTrans and the PAAC ( Figure 3- 17) for a final field operational test.
26 Source: Final Report, Integrated Collision Warning System Final Technical Report. December 2004.
27 University of California PATH and Carnegie Mellon University Robotics Institute. Integrated Collision Warning System Final Technical Report. December 2004. Federal Transit Administration 29
Assessing the Business Case for Transit IVBSS Final Report
Figure 3- 17: SamTrans and PAAC ICWS Buses28
Using knowledge from the original development projects, FCWS and SCWS were combined into a single system to provide an ICWS. The systems are linked through a mutual computer, allowing each to operate independently. The mutual computer allows the passing of critical data between the two systems, such as objects that move from the side to the front, via a serial link.
The systems are differentiated via a “ plane” that passes vertically through the front wheels of the bus. FCWS processes all objects in “ front” of the plane and SCWS processes all objects “ behind” the plane. The system does not include collision warning or object detection for the rear of the bus.
An integrated DVI was installed on the vehicles using the detection techniques for the FCWS and the SCWS. A single display ( Figure 3- 18 and Figure 3- 19) informs the operator of impending crashes at the front or at the side of the vehicle.
Figure 3- 18: ICWS DVI29
28 Source: Final Report, Integrated Collision Warning System Final Technical Report. December 2004.
29 Ibid Federal Transit Administration 30
Assessing the Business Case for Transit IVBSS Final Report
Figure 3- 19: Installed ICWS DVI30
The field operational test was completed at SamTrans and PAAC in 2003, and there is a specification for the system. However, the system is not commercially available and there are no further plans to test it.
3.4.5 Pedestrian Warning Devices
In January 2007, WMATA began a pilot program of PWD on its transit bus fleet. WMATA installed a special warning strobe atop its test fleet of Metrobuses. The yellow warning strobe light ( see Figure 3- 20) warns pedestrians and motorists of an approaching Metrobus. The strobe lights resemble the warning lights on school buses to increase vehicle visibility. As stated in WMATA’s press release, “ Metro is the first transit agency in the United States to test warning strobe lights atop buses. We believe this is another helpful safety tool designed to improve pedestrian safety throughout the region.”
Figure 3- 20: WMATA Safety Warning Strobe Light31
3.4.6 Truck Collision Avoidance Suppliers
The market for collision avoidance systems on commercial vehicles far exceeds the available transit- based systems. A prior IVBSS report identified 18 suppliers that currently supply collision avoidance and related technologies to the commercial vehicle market. Of those 18 suppliers, only two companies ( Clever Devices and Mobileye) are currently marketing their systems to the transit industry.
30 Ibid
31 Source: Washington Metropolitan Area Transit Authority Press Release, www. wmata. com Federal Transit Administration 31
Assessing the Business Case for Transit IVBSS Final Report
Three major suppliers of truck collision avoidance systems were contacted to discuss their interest in the transit collision avoidance market. Each supplier voiced similar concerns for not venturing into the transit bus market. These concerns included:
1. Market size is too small.
2. Urban setting presents challenges their technologies are not designed to address.
3. Transit operating speeds are lower than commercial vehicle speeds.
4. Driver fatigue is less of a concern in transit.
The interviewed companies stated that the market size and the perceived low demand did not warrant marketing their systems to the transit industry. The transit industry has limited funding for general purchases, with very little or nothing to spare for the research and development of collision avoidance systems. Therefore, few transit authorities are interested in using their limited funds to install a system that is not yet validated. In addition, transit buses account for just 70,000 of registered vehicles, compared to the 1.7 million commercial trucks registered in the United States32. The sheer number of commercial trucks allows a company to make a sound business case to develop a commercial system. A transit authority will only provide limited orders for a particular system. The suppliers would not reap the benefits of mass production, as each transit authority is likely to special order a system.
The systems for commercial vehicles are used in a highway environment, which is characterized by high speeds and minimal contact with non- vehicular objects. In this operating environment, it is logical for the commercial systems to label most detected objects as a potential threat. In contrast, transit buses operate in an urban environment with many inanimate objects and at a slower traveling speed. These conditions render the assignment of a threat very difficult, with an increased likelihood of false alarms. Systems with high false alarm rates will be ineffective if the operators begin to ignore all warnings. One supplier installed its collision warning system on Greyhound buses. Due to the high false alarm rate, Greyhound removed the systems after a single year in operation. The supplier has since completed numerous hardware and software revisions to address the false alarm rate; however, Greyhound has not ventured back into the collision avoidance system market.
One supplier stated that commercial truck drivers experience different driving conditions than transit bus operators. Long- haul drivers frequently experience fatigue. Driver fatigue/ inattention may lead to lane drift and smaller following distances. Each of the suppliers interviewed used CAS to assist fatigued drivers. With the shorter driving times, bus operators may not experience the same fatigue as commercial truck drivers. ( It should be noted that bus drivers might experience fatigue or boredom because of driving the same or similar routes frequently.)
One benefit of interviewing the suppliers was the in- field experience they had gathered from the operators. One supplier distributed surveys to approximately 300 drivers to solicit feedback on their experience with the collision avoidance system. Of the drivers surveyed, 75 percent felt the system provided adequate feedback to facilitate safer operation of their vehicle. Additionally, 65 percent of drivers felt that the LDWS reduced their fatigue. The drivers were more likely to pull over after receiving an alert for lane drift. The result was presumably reduced fatigue by avoiding additional driving.
An interviewee stated that driver understanding and willingness to accept the product was directly related to the management’s implementation of the system. As management provides a proactive approach to the system implementation, the drivers appear to be more receptive to integrating the system into their
32 Source: Bureau of Transportation Statistics, 2003 Federal Transit Administration 32
Assessing the Business Case for Transit IVBSS Final Report
routine. This information demonstrates that the successful incorporation of any collision avoidance system will require the support of not only the upper management but also the support of the local union.
Federal Transit Administration 33
Assessing the Business Case for Transit IVBSS Final Report
4.0 BENEFIT- COST ANALYSIS
The objective of this chapter is to conduct a benefit– cost analysis of collision avoidance systems. The analysis is preliminary, as the effectiveness of these systems in mitigating or preventing collisions ( i. e., the benefits) has yet to be fully determined through operational tests. In addition, because of limited development and deployment, the total cost to purchase and maintain these systems remains uncertain. To address these uncertainties, the study’s benefit- cost analysis has relied on a detailed assessment of current bus collision frequencies and costs; an assessment of the likely effectiveness of various existing and potential future safety systems in addressing a broad range of well- documented crash types; and current expectations of the costs of safety systems once commercially available. The results of this chapter include identification of those conditions under which the benefits of various bus collision avoidance systems— either individually or collectively— are expected to exceed the costs. In addition, this chapter discusses several sensitivity analyses performed on the results and the implications of the results for transit agencies.
Key findings of the benefit- cost and sensitivity analyses include the following:
• Collisions are a substantial source of costs to transit operators. Fatalities, injuries, and property damage constitute, on average, between 5 and 10 percent of bus operating costs. On average, U. S. agencies spend over $ 4,000 per bus in collision- related costs each year.
• Agencies face numerous non- financial collision effects, including strains on administrative human resources, negative public perceptions of bus safety, and even contention with lawmakers and funding authorities.
• Only SODS and combinations of systems containing SODS “ passed” the benefit- cost test with a ratio above one consistently ( under a range of assumptions about collision rates and technology costs).
• One other strong performer was pedestrian detection systems; however, this had a benefit- cost ratio above one only under a minority of the scenarios considered as part of the sensitivity analysis.
• While many of the data used to populate the benefit- cost analysis are reliable, some are subject to significant uncertainty along several dimensions. Thus, the results of the benefit- cost analysis should be interpreted with caution due to the subjective nature of some of the analysis inputs, and the sources of raw data used in developing estimates of the benefit- cost ratios should be considered. Data sources are indicated throughout this chapter, along with qualitative explanations of their shortcomings and, in some cases, attempts to quantify the level of uncertainty.
• Sensitivity analyses were performed to account for some of the uncertainty in the results. These sensitivity analyses provide additional insights into the overall performance of various technologies and, in general, confirm the baseline findings that only SODS and combinations containing SODS consistently achieve benefit- cost ratios above one.
Figure 4- 1 presents an overview of the computational framework for the benefit- cost analysis, which corresponds with the structure of this chapter. Section 4.1 describes the first step of the analysis framework ( including steps 1a through 1c), culminating with an estimation of IVBSS benefits, while Section 4.2 discusses the total costs of IVBSS ( step 2). Section 4.3 presents the results of the benefit- cost analysis ( step 3).
Federal Transit Administration 34
Assessing the Business Case for Transit IVBSS Final Report
In addition, Section 4.4 describes the parameters and results of a sensitivity analysis. Finally, Section 4.5 describes the implications of the benefit- cost results, particularly as they relate to agencies with insurance coverage for accidents and liability claims.
Figure 4- 1: IVBSS Benefit- Cost Computational Framework 1aCollision frequencies1bCollision costs1cCollision prevention rates1Benefits of IVBSS deploymentxx= 2Costs of IVBSS deployment3Benefit- cost results÷Data from NTD and 6 participating agenciesDetailed collision data from 2 agencies1aCollision agencies
4.1 Benefits of IVBSS Deployment ( step 1)
IVBSS for transit buses offer a wide range of potential benefits to agencies and to society in general. These benefits include the following:
• Direct cost savings to agencies investing in IVBSS attained through avoidance or mitigation of collisions
• Improved public image
• Reduction in damage and injury claims related to non- collision events ( e. g., passenger falls on board) due to improvements in operator training and safety practices enabled by IVBSS
• Reduction in external costs indirectly related to collisions, such as congestion
• Improved ridership and customer satisfaction
• Improved training capabilities
This analysis considers only the first of the above benefit categories— direct cost savings to investing agencies attained through avoidance or mitigation of collisions. As a result, the analysis excludes qualitative agency benefits and all social benefits.
Estimating the financial benefits stemming from collision avoidance requires the following steps:
• In step 1a, the frequencies of collisions are quantified according to a matrix of collision types and severities.
• In step 1b, the costs of collisions are quantified according to the matrix of collision types and severities.
• In step 1c, available data are analyzed in order to estimate collision prevention rates. Federal Transit Administration 35
Assessing the Business Case for Transit IVBSS Final Report
• Finally, these three pieces are assembled to produce summary estimates of IVBSS collision avoidance benefits.
Data Sources
There are two primary sources of national transit bus collision data: 1) the National Transit Database ( NTD) and 2) the Buses Involved in Fatal Accidents ( BIFA) database. 33 NTD is an FTA- maintained database covering operational characteristics, service characteristics, capital assets, revenues, and financial performance of the more than 600 transit agencies receiving Section 5307 federal formula funds. NTD’s Safety and Security module contains data on major and non- major collisions, defined by the FTA as follows:
• Major collisions are those that involve at least one fatality, at least two injuries, and/ or property damage exceeding $ 25,000 ( including damage sustained both by the transit agency and by third parties).
• Non- major collisions are those in which at least one person was injured and/ or total property damage exceeded $ 7,500. ( Agencies must also report non- major collisions.)
NTD data corresponding to these reporting thresholds are available for calendar years 2002 through 2005. Prior to 2002, NTD used a lower reporting threshold, reporting all accidents with property damage in excess of $ 2,000.
BIFA is a census of all fatal bus- involved collisions, derived from the Fatality Analysis Reporting System ( FARS) and maintained by the University of Michigan Transportation Research Institute ( UMTRI). UMTRI also performs detailed follow- up investigations with operators, witnesses, and transit agencies for each bus collision. Fewer than 100 fatal transit bus collisions are recorded each year; furthermore, detailed BIFA records are unavailable to the public. As a result, this data source was not used in the estimation of collision avoidance benefits.
Prior to this study, relatively few data had been collected on the cost and frequency of “ minor” collisions ( non- injury collisions with property damage below $ 7,500). To develop a more complete picture of all bus collision types ( of minor collisions in particular), FTA requested and received internal collision records from six transit agencies (“ participating agencies”). The participating agencies have a combined active fleet of 7,000 buses, or approximately 10 percent of the national transit bus fleet. Collision records covered an average of almost 3 years of data per participating agency and yielded 17 “ agency- years” of data in total. Although no standardized methods of collecting collision data exist across transit agencies, the collision records provided sufficient information to support the segmentation of collision types and severities developed for this study. Agency data also provided cost and frequency data on minor collisions not reported to NTD.
Step 1a: Collision Frequencies
The effectiveness of bus collision avoidance systems will vary depending on collision type and collision severity. For example, safety systems are unlikely to prevent or mitigate the severity of high- speed, right- angle (“ T- bone”) collisions, but might be effective in reducing the frequency and severity of sideswipe collisions. Given these expected differences in system effectiveness based on each collision’s characteristics, the benefit- cost analysis reflects the anticipated differences in investment benefits for each collision type and level of severity ( note that prior benefit- cost analyses of safety systems have not utilized a detailed segmentation of collision types and severities). This subsection presents the
33 The General Estimates System ( GES) is a commonly used source for national crash data. However, since GES includes only sample crash data, it does not accurately reflect transit bus- involved collisions. As a result, GES data are not used in this benefit- cost analysis.
Federal Transit Administration 36
Assessing the Business Case for Transit IVBSS Final Report
segmentation of collision types and severities used by the benefit- cost analysis and the estimates used to populate that segmentation.
Collisions can be categorized in many ways. However, the categorization scheme selected for this analysis was based on the organization of collision data from NTD and the participating agencies in order to enable usage of a broadly representative data set. Bus collisions were segmented into a matrix of five collision severities and seven collision types, yielding 35 unique severity- type combinations. The five collision severity categories include the following:
• Major Fatal: Collisions reported to NTD as major collisions with at least one fatality
• Major Non- Fatal: Collisions reported to NTD as major collisions with no fatalities ( i. e., at least two injuries and/ or property damage in excess of $ 25,000)
• Non- Major with Injury: Collisions reported to NTD as non- major collisions with at least one injury
• Non- Major PDO: Collisions reported to NTD as non- major collisions with no injuries ( i. e., property damage between $ 7,500 and $ 25,000)
• Not FTA- Reported: All other collisions not reported to NTD ( i. e., no injuries, property damage below $ 7,500, and recorded internally by the transit agency as a collision) 34
The seven collision type categories include:
• Frontal: Frontal collision with another vehicle ( e. g., bus rear- ends a vehicle)
• Rear: Rear collision with another vehicle ( e. g., another vehicle rear- ends bus or bus rolls backward into another vehicle)
• Sideswipe: Sideswipe with another vehicle ( e. g., bus scrapes the side of another vehicle during a right turn maneuver or while changing lanes)
• Angle: