Design And Implementation Of An Early Warning Traffic System Based On A Stream Data Model Of Topology

With the rapid development of urban construction,the traffic safety situation becomes more and more serious on the urban expressway.In order to reduce the number of traffic accidents,the issue for establishing an effective early warning mechanism needs to be resolved in the study of traffic safety.It is found that the relation between the level of traffic safety and the traffic state hides in the stream data from various sensors.Therefore,it's an effective way for improving the level of traffic safety to mine the available decisional rules timely from the big data of traffic.Under this situation,the research mainly focuses on three aspects:(1)Aim at dealing with the real-time,huge,and dynamic stream data of traffic,we propose a suit of real-time data processing framework which is based on the data stream computing platform ‘storm',the distributed message system ‘Kafka',and the document type database ‘MongoDB'.(2)We give an improved algorithm,i.e.,fuzzy C-means(FCM)based on the Topology model in this paper.A simulating calculation,performed by software VISSIM,confirms the effectiveness of the proposed algorithm for recognizing and grasping the traffic state which has a distinct time and space variant characteristics.(3)In order to extract decisional rules,we design and implement a traffic early warning system based on the proposed algorithm.By testing the traffic flow dataset from the sensors on the second ring expressway in Beijing,the system is verified to be effective and feasible.So,we believe that the system has a practical and promotional value.