CHAPTER 3

Survey Results

This chapter includes the results of the survey administered to American transit agencies and commuter rail agencies. The purpose of the survey was to understand and document how these agencies are addressing electronic surveillance for collision avoidance at railroad-highway crossings.

Survey Design and Analysis

The survey was designed in two parts. The first part, comprised of eight questions, sought perspectives on the monitoring and surveillance of rail transit and commuter rail crossings from a wide range of individuals in academia, consultants, vendors, and a broad range of transit agencies. The second and remaining part of the survey was focused on installed systems and experience with such systems. The questions were framed to get information about data collection and analysis methodologies (e.g., crash frequencies, unintended events, and trespassing behaviors), decision criteria, application of data, system components and configuration, capital and operating costs, funding sources, measures of effectiveness, and challenges and constraints (legal, institutional, and technological). The survey had 23 questions and the complete survey is provided in Appendix A.

Survey Respondents

The survey respondents were professionals from rail transit and commuter rail agencies, academic researchers, consultants, vendors, professionals from state transportation agencies, and others. A total of 45 survey responses were received, with 35 complete responses. The respondents came from a variety of rail transit and commuter rail agencies. Not all had direct experience with electronic surveillance of rail crossings. The other responses were from academic researchers, consultants, vendors, oversight agencies, and retired rail transit agency professionals. Appendix B provides a listing of agencies and individuals who participated in surveys and/or interviews that helped develop case examples. Some respondents did not include transit agency/affiliation or title information and are not listed in Appendix B. Summaries of survey responses are shown in Appendix C; not all respondents answered all questions. The survey results are discussed further in this chapter in subsequent sections.

Reasons for Using Electronic Surveillance

Figure 5 shows various reasons agencies are using electronic surveillance of rail transit or commuter rail crossings. Avoiding train collisions with vehicles and pedestrians was chosen by most respondents [over 80% (36)].

Systems and Technologies

There are various systems in use from a basic camera set up for photo enforcement (for example, used at rail crossings in corridors Metra and LACMTA operate) to CCTV and LIDAR-/RADAR-based system (as used by Network Rail). Some systems comprise of fiber infrastructure, switches, National Electrical Manufacturers Association (NEMA) Box, and pan-tilt-zoom (PTZ) or Fixed Camera. Some grade crossing monitors grade crossing operations and generate alerts for abnormal behavior for train operators and enforcement agencies. Yet another system is set up for the detection of trespassers, collisions, vehicle dwelling, blocked crossings, fires, equipment failures, timing, and other data.

The more advanced system has an AI interface that automatically detects grade crossing violations, traffic, trains, and signal activations. The hardware on the edge consists of networking equipment and a modem. Cloud services include AWS EC2 Instances and Continued Internet 4G service.

Some monitoring systems have been expanded to observe operational performance in addition to safety violations. In the TRAINFO system trains are detected with acoustic sensors; video cameras can be added if visual features are needed. The data are wirelessly sent to the cloud for analysis. Sound signatures are developed to determine train movement characteristics and predict when crossings will be blocked and cleared. Traffic data can be integrated to predict travel time delays for motorists up to 10 minutes before the train arrives. Application programming interfaces are used to integrate this information into roadside signs, traffic management centers, computer-aided dispatch software, and mobile apps.

Decision Criteria

Figure 6 shows different decision criteria rail transit and commuter rail agencies use in implementing electronic surveillance at rail crossings. Safety metrics such as crash frequency, near misses, and so on were chosen by over 80% (36) of the respondents as a decision criterion. Over 70% (32) of respondents chose “prevent trespassing and suicides – crashes and injuries saved” as the key criterion.

Applications

Based on the literature review, selected documented experience, and anecdotal evidence, a list of applications was compiled and provided as choices for the survey question about reasons for electronic surveillance, as shown in Figure 5.

The respondents who had experience with installed electronic surveillance provided insights about the nature and scope of applications. The motivation for electronic surveillance varied from agency to agency. Vehicles, pedestrians, and trespassers and their behavior and violations were monitored. There was interest in understanding the effectiveness of treatments through before and after videos, so the system was moved from one crossing to another that expanded the utility of the system. Along the same lines, the setup was used by some to communicate the vital health status of the traffic signal, the status of the gates, island relay, and the direction of the traveling train. Thus, there was less use of conductors for the interconnection. In some cases, the major intent was photo enforcement that was used to capture left turn violations and issuance of citations. This was particularly the case with the Metra and LACMTA. Some used the system for research, real-time hazard detection, and maintenance support. In some cases, the system has been used to identify types of violations to formulate better engineering and enforcement solutions. The system is being used to evaluate the effectiveness of applied engineering solutions. For example, the system considered the

effectiveness of quad gates at a particular crossing. In the case of Network Rail, obstacle detection (road vehicle, pedestrian, etc.) is interlocked with the train control system.

From an operations perspective, TRAINFO’s applications include traveler information systems, traffic management, emergency services, and analytics. TRAINFO notifies road users about blocked crossings, expected delays, and re-route options through various methods such as roadside signs and mobile apps. For assistance with traffic management purposes, information is integrated into traffic signal management systems to adjust signal timing before, during, and after a blockage event to reduce traffic congestion; it can also be used as an alternative to signal pre-emption. For assistance to emergency services, information is integrated into tactical maps and computer-aided dispatch software for call-takers and dispatchers to help first responders re-route or select units that will not be impacted by a train. For analytics, TRAINFO offers an online data portal with interactive graphs that provide blockage trends and detailed statistics (e.g., blockages by time, duration, train speed, and train length). TRAINFO also produces risk models for first responders to identify which crossings impact emergency calls the most and the magnitude of these applications in the research and development phase include connected and automated vehicle information and integration into Apple CarPlay and Android Auto.

Barriers, Challenges, and Constraints

Figure 7 presents different barriers, challenges, and constraints faced by rail transit and commuter rail agencies when deciding on or maintaining electronic surveillance of grade crossings. Most respondents [over 70% (33)] considered funding as the key barrier, and over 60% of respondents rated cost as a major barrier. Other important barriers were complexity, technological, and legal in nature. Institutional support was not considered as big a barrier by respondents.

Effectiveness

The measures of effectiveness that survey respondents were asked about were the

• Rate of compliance/violation;
• Number of collisions avoided;
• Number of trespassers avoided;

• Number of instances and duration of blockage on surface streets; and
• Crash frequencies.

Figure 8 presents different measures of effectiveness chosen by respondents. Safety is the dominant reason for electronic surveillance of rail crossings. Over 60% (30) of respondents chose the rate of compliance/violation, the number of collisions avoided, the number of trespassers avoided, and crash frequencies as key measures of effectiveness.

Success Factors

Figure 9 provides insights regarding success factors. Most respondents [67% (30)] considered the availability of funding as critical in establishing and using an electronic surveillance system. Similarly, about 64% (29) considered data and analysis possible because such systems are helpful for decision-making and make it a worthwhile and successful investment. Also, 55% (25) thought the system could be successful if it is maintained and updated well. Video analytics is of major interest to respondents. Other factors were also chosen as important for success but did not rank as highly as others mentioned.

Causes for Failure

Figure 10 identifies causes for failure. Over 60% of respondents identified a lack of institutional support and the system not being maintained or updated as reasons for failure. A lack of technical resources was also considered as an important reason for failure.

Cost

The setup cost varied. In some instances, especially when the right-of-way is not owned by the agency, set-up costs can be high to secure approval and deal with labor costs. In other instances, as is the case with most light rail systems, since the right-of-way and system is owned and operated

by the same entity, it is much cheaper to develop and establish the infrastructure needed to install the system. The cost ranged from less than $10,000 per location to over $100,000 per location. This cost difference is also because some systems, like Network Rail, are quite complex and sophisticated and use both CCTV and LIDAR-/RADAR-based systems to monitor rail crossings and provide appropriate alerts.

The recurring cost also varied per location, but most indicated it to be less than $10,000 per location. Most of the recurring cost is related to data collection, storage, and processing. Some advanced systems have demonstrated great success in video analytics and the resulting information for decision-making.

Funding

Over 88% of the respondents indicated that federal funding is the key source for the establishment of electronic surveillance systems at rail crossings. State and local funding has also allowed others to sustain the installed system. Most stable funding from a mix of federal, state, and local sources has been obtained by UTA Transit and TriMet agencies. In many instances, support has come to establish and maintain photo enforcement, particularly in the Metra and LA Metro cases as seen in Figure 11.