A Pedestrian Crossing Warning System in Cleveland that Included 24 Buses and Three Instrumented Intersections Reduced Bus Driver Reaction Time to Pedestrian Conflicts by 19 Percent.

An Enhanced Transit Safety Retrofit Package (E-TRP) Was Deployed and Evaluated in the Greater Cleveland Area.

Date Posted

Connected Vehicle (CV) Infrastructure - Urban Bus Operational Safety Platform

Summary Information

Federal Transit Administration (FTA) developed an enhanced version of the Transit Safety Retrofit Package (TRP) system that was originally part of the United States Department of Transportation (USDOT) Safety Pilot Model Deployment, a large-scale Connected Vehicle (CV) deployment. The enhanced TRP (E-TRP) was based on the earlier system, focusing on reducing pedestrian and vehicle conflicts with transit buses in the greater Cleveland, Ohio, metropolitan area.

Key technologies included Dedicated Short-Range Communications (DSRC) for  vehicle-to-vehicle and vehicle-to-infrastructure communication, High-precision Global Navigation Satellite System (GNSS) for vehicle tracking, and Forward Looking Infrared  (FLIR) cameras for enhanced pedestrian detection.

The deployment included two subsystems: (i) In-Vehicle Subsystem (IVS), a transit vehicle-based subsystem, and (ii) a Roadside Subsystem (RS) at each of the selected street intersections. The E-TRP featured enhanced versions of the Pedestrian in Crossing Warning (PCW) and Vehicle Turning Right in Front of Bus Warning (VTRW) CV applications.

A total of 24 Greater Cleveland Regional Transit Authority (GCRTA) transit buses were equipped to enable field testing at three locations: one signalized intersection, one non-signalized intersection, and one mid-block crossing.


The buses operated in revenue service for a period of six months, from February to August in 2018. During the first month, the IVS was put into “cloaked” mode (i.e., the system did not display alerts to drivers). The data for those concealed alerts and the drivers’ reaction time towards engaging the brakes without knowledge of the alerts were recorded. In the remaining five months, the system was taken out of cloaked mode, and the drivers received live alerts when traversing through the equipped sites.

Evaluation analysis areas included (1) System Performance, (2) Safety Impact, (3) Return on Investment, and (4) Driver Acceptance. The performance measures used in each analysis area were (1) false alarm rate , (2) collision reduction, (3) cost-savings, and (4) usability, perceived safety benefits, unintended consequences, and desirability, respectively.

The estimated cost savings due to collision reductions was determined with by using the USDOT’s Value of a Statistical Life (VSL) method. Note that the evaluation excluded the E-VTRW application due to the lack of the presence of dedicated short range communication (DSRC)-equipped vehicles (e.g., light vehicles). In addition, only 13 out of 751 E-TRP drivers participated in the survey to assess driver acceptance of the application. Therefore, driver acceptance results were not statistically significant and may not be representative of the system. 


  • Quantitative data analysis of driver braking within 2.5 seconds of warning alerts issued showed that the application increased driver response by 17 percent, from 12 percent to 14 percent compliance. Furthermore, the reaction time average of the driver was reduced by 19 percent from 1.6 seconds to 1.3 seconds,
  • Of alerts issued by the system, 81 percent were correct alerts, while 10 percent of alerts were incorrect, and 9 percent were false alarms. False alarms were caused primarily by lighting and environmental conditions.
  • The return on investment analysis showed an estimated average annual system benefit of $106,452. The total cost to deploy the system across the larger transit network was estimated at $2,163,180. These findings suggested that the system implementation cost could be recuperated in 244 months (more than 20 years) as a result of risk reduction. However, with the development of the technology in the future, the system costs were anticipated to drop substantially and would shorten the period of return on investment.
Results Type
Deployment Locations