Real time evaluation of shortest remaining processing time based schedulers for traffic congestion control using wireless sensor networks

Pre-timed traffic signals are inefficient in optimizing the traffic flow throughout the day, resulting in greater waiting times at the intersections particularly in congested urban areas during peak hours. Traffic actuated signals use real time traffic data obtained from sensors at the intersections to service queues intelligently. We developed a test bed for the real time evaluation of adaptive traffic light control algorithms using the microscopic traffic simulation open source software, SUMO (Simulation of Urban Mobility), and the AVR 32-bit microcontroller. An interface was developed between SUMO and the AVR microcontroller in which we used the simulation data generated by SUMO as an input to the microcontroller which executed the scheduling algorithms and sent commands back to SUMO for changing the states of the traffic signals accordingly. We implemented four scheduling algorithms in SUMO through the AVR microcontroller, the effect of the algorithms on the traffic network was studied using SUMO and execution times of the scheduling algorithms were measured using the AVR microcontroller. Through this interface, scheduling algorithms can be evaluated more effectively and accurately as compared to the case in which the algorithms are fed with data using pseudo random number generators.