A novel algorithm for optimized, safety-oriented dynamic speed regulation on highways: INCA

This paper reports on implementation of a novel algorithm for optimizing the dynamic regulation of motorway traffic using variable speed/information displays, with the aims of increasing traffic safety and reducing unnecessary delays. The new control system incorporates an adaptive data and information fusion approach based on optimization of an objective function. This data driven approach takes historical data and, using the objective function, extracts knowledge about the specific reliability of each algorithm under varying traffic conditions. The extracted knowledge is integrated into a flexible framework incorporating a logit probability model. The trained system produces a control decision for each display, taking into consideration neighboring detectors. The implemented objective function can be used for assessing, optimizing, and monitoring the performance of line control system as a whole as well as of individual incident detection and state estimation algorithms. The system has been implemented and installed in a Munich traffic management center. Following calibration on 2 months of data, the system was tested for 2 months on-line in open-loop mode. During this test, the optimized LCS framework out-performed the previously implemented control according to several measures of quality, such as detection/false alarm rates and the objective function.