Taxi Abnormal Operation Detection Technology Based On GPS Data

The intelligent transportation system(ITS)provides convenient,safe and efficient city services,such as electronic maps,travel planning and resource scheduling.The massive data collected by sensors can be used for traffic flow analysis,path planning,anomaly detection,passenger flow analysis and a series of studies.In many applications,real-time anomaly detection of pavement vehicles is an important way to ensure normal traffic and passengers’ safety in urban areas.However,intensive road networks and complex environments make existing anomaly detection algorithms less precise or higher false alarm rates.Therefore,using vehicle sensors to study an accurate and efficient vehicle abnormal operation detection method has important significance and application value for the protection of urban traffic safety and passengers’ interests.This paper mainly studies the related technologies involved in the taxi abnormal operation detection process.The main contents are as follows:(1)Aiming at the problem of diversification of vehicle GPS data in complex urban environments,the survival analysis is applied to the field of missing trajectories,and the survival analysis based on anomaly data of missing taxi GPS trajectories is proposed.The experiment results prove that the survival analysis method can effectively classify the missing events of GPS trajectories in the city and adopt different complement methods according to the missing of different trajectories,which is beneficial to improve the efficiency and accuracy of trajectory complementation.(2)Aiming at the problem of high false alarm rate in the existing abnormal trajectory detection algorithms,this paper studies the idea of isolation,designs a spatial and temporal isolation-based online anomalous trajectory detection(ST-IBOAT)algorithm.The input trajectories are used to judge the anomalous degrees,and the abnormal sub-trajectories and anomalous degrees are fed back in real time.The experimental results show that the anomaly detection using the double window can effectively improve the detection accuracy with the accuracy up to 99%.(3)On the basis of the above research,the core function modules are coded and the abnormal operation detection system of taxi is designed and initially realized.The system is tested using a real taxi GPS data set.The results show that the taxi abnormal operation detection method designed in this paper can effectively solve the problem of false alarm rates in complex urban environments,and can feedback abnormal sub-trajectories and anomalous degrees in real time.