Development of an Adaptive Control Algorithm for Arterial Signal Control

An optimization algorithm was developed to optimize the signal offsets along arterials in real-time based on volume data from loop detectors. The algorithm performance was tested with the VISSIM microscopic model on three real-world arterials. The algorithm generated offsets had better performance than the offsets generated using the widely used SYNCHRO optimization tool. The real-time update of offsets generated by the algorithm was also evaluated using a custom made simulation testbed. The results indicate modest improvements in traffic performance over the SYNCHRO timings. A data filtering algorithm was developed to estimate the systematic error of detector data. Application of the algorithm at a real world intersection with multiple detectors resulted in significant error reduction on high volume approaches. A clustering algorithm was also applied to the intersection detector data to determine both the traffic demands for developing TOD signal timing plans, and the associated plan switching times. The new timing plans outperformed the conventional TOD plans based on traffic volumes obtained at predetermined time periods.