A Real-Time Processing System for Massive Traffic Sensor Data

With the continuous expansion of the scope of the transportation sensor networks, a new kind of data, namely traffic sensor data, becomes widely available. Traffic sensor data gathered by large amounts of transportation sensors shows the massive, continuous, streaming and probabilistic characteristics compared to traditional data. In order to satisfy the requirements of different traffic sensor data applications, the capability of real-time processing for massive traffic sensor data is emergently needed. In this paper, a Real-Time Processing System (shorted as RTPS), which adopts the decentralized distributed architecture to support the parallel processing of traffic sensor data, is presented with a case study of a real world application about vehicle license plate recognition data. And the parallel computing model behind RTPS and corresponding programing interface are proposed. The experiment based on application of vehicle license plate recognition data shows that our system has good scalability and the processing performance increases in linear progression as the number of processing nodes increases.