Designing and Developing a Model for Detection of Unusual Traffic Condition at Intelligent Signalized Intersection Equipped with SCATS

Today, one of the most significant points of interference in streets is the signalized intersections; therefore, solving problems of traffic at intersections can increase the capacity of urban transportation. Inability to diagnose the traffic conditions results in the lack of proper timing, phasing and cycle length, all of which are attributed to the abnormal factors concerning intelligent control systems. In this paper, in addition to the introduction of abnormal traffic conditions at signalized intersections, an attempt has been made to intelligently diagnose anomalies for both an approach and its entire intersection. For this purpose, by making use of the data based on the GPS of users' cell phones extracted from NESHAN Application, which consists of 10-minute average speed in streets ending to an intersection, and by behavioral matching with the data concerning the volume and saturation rate in SCATS, and meantime, by analyzing the fundamental traffic relations, an attempt has been made to diagnose the abnormal traffic conditions through SCATS at Toos-Danesh intersection of Mashhad, in which abnormal conditions including the detecting of heavy traffic conditions when the traffic is light and vice versa. To achieve more accuracy, the method was built based on both quartiles and percentiles of DS, degree of saturation, and ADS, average degree of saturation, in SCATS. Finally, anomaly detection based on the 10th and 90th percentiles had 100 percent accuracy and the one based on the 1st and 3rd quartiles had between 57 to 80 percent accuracy, which have been checked by two real datasets.