1375210 - Evaluating the Effectiveness of Vehicle-Mounted Variable Message Signs (VMS) in Protecting Roadside Incident and Service Personnel: A Deep Learning Study

This study aims to evaluate the Variable Message Signs (VMS) mounted over the cab of emergency vehicles as a safety countermeasure used to aid in protecting roadside incident and service personnel. The team worked with the local Safety Service Patrol (SSP) program to collect video data from their emergency service vehicles. The data are recordings of traffic environments in which emergency vehicle operators were providing road service on roadside. The data were collected under naturalistic conditions. Over 50 hours of videos were collected and a total of 40 different ASAP vehicle stops at various locations were observed. This study was interested in how motorists would react to the situation with the presence of an ASAP vehicle on the roadside and with or without the use of the vehicle-mounted VMS board. Deep learning was utilized to detect and track vehicles in videos. A total of 11,338 vehicles passing the ASAP vehicle (on roadside) were detected in 135,946 frames of videos. Vehicle trajectories were extracted from the videos. With the trajectory data, the vehicle movements such as passing speeds and lane positions are explored. Three statistical models are built to establish the relationships between factors and to examine the impact of VMS on passing vehicles as an indication of the effectiveness of VMS as a countermeasure to protect roadside incident and service personnel. The explanatory variables include the use of VMS along with other factors such as vehicle class, time of day, weather condition, roadway geometry, and traffic density. To capture the potentially unobserved factors for the features of stop locations, this study developed Weighted Random Intercept Models to estimate the relationships between factors. Modeling results showed the relationships between the use of VMS and the passing vehicles’ movements measured by speed, lane change, and distance to ASAP vehicle when moving over. When the VMS was used, drivers were more likely to move over (change lanes) and slow down than in the cases when the VMS was not used. The odds for a vehicle to move over is 95 percent higher when the VMS was used. The findings imply that the use of VMS can have a positive impact on traffic especially passenger vehicles. Therefore, vehicle-mounted VMS can be considered as an effective countermeasure to protect roadside incident and service personnel.