Research Of Traffic Incident Detection Algorithm Based On Laser Rangefinder And GPS Information Fusion

In order to timely detect the traffic congestion incidents of city road and reduce travel delays, property damage and casualties, it is very important to study the fast, efficient and accurate algorithm to automatically detect traffic incidents(AID, Automatic Incident Detection).The source of data is single because it is collected by the stationary vehicle inspection for most AID algorithms. On the one hand, most AID algorithms are developed for the highway less in urban roads. On the other hand, they are mainly used to detect whether the incident occurred rather than the development trend of congestion event.In view of the above problems, two new AID algorithms based on laser rangefinder and GPS information fusion are proposed, combining with the new generation data acquisition technology and artificial intelligence theory.Firstly, the characteristics of city road traffic flow and the principle of traffic incident detection are introduced. On this basis, a new AID system is proposed based on the analysis of the principle of laser rangefinder and GPS; Secondly, it is pointed out that there are some problems if most AID algorithms applied to the detection of city road traffic congestion in our country. Moreover, a new information fusion model is proposed based on the analysis of the principle of laser rangefinder and GPS information fusion, which provides a reference for the selection of the feature vectors of the classifier; Finally, a new SVM-AID algorithm is proposed for the two states of the city road traffic congestion, and a new PSO neural network AID algorithm is proposed for the four states: traffic normal, traffic congestion, congestion intensifies and congestion mitigation. The feature vector, the design idea and working procedure are constructed for the two new AID algorithms. In the end, the validity of the new AID algorithms are tested by the measured data.Test results show that compared with the classical AID algorithm, the new algorithms have a higher detection rate, a smaller false alarm rate and a shorter average detection time.