Dynamic Detection of Multi-forked Tree Gantry Anomaly Based on Isolated Forest Algorithm

The large-scale deployment and networking of the expressway electronic no parking toll collection (Electronic Toll Collection, ETC) system has greatly improved the efficiency of vehicle traffic. However, the failure of the gantry equipment will seriously affect the stable and reliable operation of the ETC system, and will also bring great work burden to the daily operation and maintenance personnel. To address this problem, the paper proposes a gantry anomaly dynamic recognition algorithm, DTS-BFS. The algorithm decomposed the gantry topology map structure into the vehicle dynamic trajectory, for the first time, the abnormal detection, and according to the time series characteristics of the gantry transaction data, can extract the abnormal characteristics of vehicle trajectory dynamic change from static network topology, and then use the isolated forest algorithm to detect gantry anomaly. The experimental results show that the algorithm can improve the accuracy of anomaly detection, and the efficiency of anomaly detection is better than other retrieval algorithms of multi-fork tree, which can provide new ideas and theoretical basis for the operation and maintenance of highway gantry.