Decision-Making System for Freeway Incident Response Using Sequential Hypothesis Testing Methods

Abstract

Recent research in advanced traffic management systems has emphasized incident detection and response to mitigate nonrecurring congestion. Existing incident response decision-making algorithms do not account for the expected losses associated with false alarms, undetected incidents, and delayed incident response. A freeway incident response decision-making system based on sequential hypothesis testing techniques is presented. The primary feature of this decision-making system is that it minimizes the sum of the expected losses associated with false response, nonresponse, and delayed responses to incidents through a dynamic programming algorithm. The results of simulation tests indicate that this algorithm performs better than typical Bayesian incident response algorithms for mean response time, false response rate, and nonresponse rate.