Research And Design Of Real-Time Streaming Trajectory Similarity Algorithm Based On Storm

"Internet +" promotes the combination of Internet technology and modern industries.All sorts of behaviors in real life can be transformed into dynamic data streams by emerging IT technology.The traditional industry begins to focus on real-time dynamic data continuously generated.The ability to extract knowledge and information from dynamic data in real time becomes more important.In the field of trajectory research,the study of real-time trajectory similarity is the foundation of many important applications.For example,real-time dynamic planning of urban road traffic and real-time arrest of criminal suspects,vehicles and so on.Therefore,it is very important to study the real-time trajectory similarity algorithm based on Storm.Considering the real-time,persistent and unlimited growth characteristics of real-time trajectory data,the traditional trajectory similarity measurement algorithm cannot meet the requirement of real-time data calculation.This thesis uses the minimum bounding sector algorithm to split the real-time trajectory into segments and designs the PDTW-LCSS adaptive algorithm based on PDTW algorithm and LCSS algorithm.The problem of trajectory similarity measurement caused by the influence factors such as abnormal noise point and different sampling rate has been solved.The accuracy of the trajectory similarity measure at the threshold edge also has been improved.This thesis aims at the design of PDTW-LCSS adaptive algorithm and implements the real-time trajectory similarity measurement system based on Storm.Based on the analysis of trajectory similarity,the Storm real-time computing framework and Java technology have been adopted as system support.Then,the architecture of the system has been analyzed and constructed from the requirement of real-time trajectory similarity measurement and a prototype system has been designed and implemented.Finally,Zookeeper has been used to build the cluster and coordinate the Storm work nodes.The similarity measure system has been tested and analyzed.The experimental results show PDTW-LCSS adaptive algorithm has good accuracy.