Urban Road Traffic Incident Auto-Detecting Based on Decision Fusion

Abstract

The objective of this study is to improve the performance of traffic incident detection algorithms on urban roads. The concept of algorithm performance reliability is introduced to make decision fusion which combines the results of the automatic incident detection algorithm based on floating car data and inductive loop detector data. The decision fusion algorithm in this article includes three modules: 1) Detection algorithm module based on inductive loop detector data; 2) Detection algorithm module based on floating car data; 3) Module of decision fusion, introduce the concept of algorithm reliability, calculate the weights of module, and use the weighted average method to make decision fusion. Finally, VISSIM simulation system was used to get the traffic flow data, and implement the algorithm using MATLAB. The simulation results show the algorithm of decision fusion is better than automatic incident detection algorithm of single data-source.