Evaluation Of Vehicle Detection Systems For Traffic Signal System Operations PDF Download

Advances in vehicle detection systems and signal controllers have produced actuated traffic signals that react to the presence of vehicles as well as adaptive traffic signal systems that predict vehicle arrivals to proactively adjust signal timings. However, one thing that has not changed over the years is the reliance on point-based data produced by loop activations or by alternative vehicle detection systems that emulate the behavior of loop detectors. Even when limited, point-based data remains the standard used by traffic engineers to measure or estimate volume, speed, density, occupancy, delay, and emissions among other performance measures. Microwave radar-based vehicle detection is one commercially-available technology used as an alternative for loop detectors. The IntersectorTM is a radar-based vehicle detection system that continuously monitors the trajectory of vehicles within the range of the device while approaching an intersection. However, trajectory information which is composed of speed, position, and vehicle length is only used to communicate the presence of vehicles over a virtual loop to the signal controller. In other words, all trajectory information is constantly discarded. If collected, discarded trajectory data can be used to directly monitor performance measures by analyzing the full path followed by a vehicle and the corresponding speed profile. A software-based data collection system developed as part of this research capable of monitoring and logging vehicle trajectories which are constantly discarded by the IntersectorTM is presented. Data collected from the system was used to directly monitor operational and safety performance measures. Data analysis procedures were developed to obtain turning movement counts, lane classifications, stopped delay, post-encroachment time, red light running statistics, and operational speeds. All procedures were written and implemented in the R programming language thus making the implementation of the procedures using low-cost hardware platforms intended for the internet of things (IOT) a possibility. The research presented represents a transformational shift in how performance measures are monitored at signalized intersections.