Research On Vehicle Motion Trajectory Patterns Based On Environmental Characteristics

The rapid development of GPS technology and the widespread use of positioning equipment have made it easier for us to obtain vehicle trajectory data.Using these data,we can reconstruct the vehicle’s trajectory and discover the vehicle’s motion pattern.The introduction of environmental features into the analysis of vehicle movement trajectories can more accurately dig out what kind of trajectory paths a moving object will choose under what circumstances.As an important data resource,vehicle trajectory data plays an important role in hotspot area discovery,road network construction,urban planning,and smart cities.It has attracted many researchers to devote themselves to the analysis of vehicle trajectories.This paper mainly completes the following three aspects of the work:1.Facusing on the frequent path problem about environmental features and uncertain trajectory data,we devise a novel algorithm named UETFP-PrefixSpan to mine frequent moving trajectory pattern from environmental features and uncertain data with strict time interval constraints.by setting the class label to distinguish uncertain trajectory data acquisition under different environmental conditions,The frequent itemsets are redefined by using probability support.This algorithm reduced the scale of projected databases and the time of scanning projected databases through reducing scanning of certain specific sequential patterns production.In this way,algorithm efficiency could be raised up.The tests results drawn that the improved UETFP-PrefixSpan algorithm has the more realistic on mining result and better efficiency.2.In view of the fact that existing prediction algorithms rarely consider environmental factors and the prediction accuracy is not high,an uncertain trajectory data prediction algorithm(EGTP)based on environmental constraints is proposed.Firstly,obtain the environmental information of the historical trajectory data,construct a new trajectory reference point by using environmental information and trajectory data,simulate vehicle uncertain trajectory data with environmental information,and then use the Gaussian mixture model to calculate the trajectory reference point data and history.The trajectory data is trained;finally,on the basis of training,the trajectory reference point and historical trajectory data are used to predict the vehicle trajectory in real time.Taking into account the environmental factors of travel,the forecast results are more in line with the actual situation.The experiment verifies that the algorithm improves the real-time performance and prediction accuracy compared with other algorithms.3.The problem of abnormal trajectory detection for uncertain vehicle trajectories with environmental characteristics.An abnormal trajectory detection algorithm suitable for uncertain trajectory is proposed.The vehicle position and travel time are taken as the spatial and temporal characteristics of the taxi.Time anomalies,spatial anomalies and spatio-temporal anomalies are divided according to the deviation of these characteristics..The trajectory data set with the same starting point is extracted from the historical trajectory data,and the entire trajectory is divided into sub track segments to calculate the similarity between the trajectories.Based on distance and density clustering,frequent and sparse trajectories are preliminarily separated on spatial features.The separation thresholds of temporal feature anomalies are determined according to the kσ criterion of data anomaly determination,and temporal characteristics are again divided,and the abnormal trajectory detection of taxis is finally achieved.Experimental results show that this method improves the clustering effect and improves the effectiveness of the abnormal trajectory detection algorithm.