At present, the usage of wireless sensor network (WSN) for traffic flow recognizing andmeasuring is a new technique on data acquisition and integration. On the basis of theexperiments used to confirm the adaptability of wireless sensor nodes, this paper proposes asystem for traffic flow detection based on WSN. First of all, the nodes acquired three kinds ofphysical characteristic parameters of sound, light, and geomagnetic field (both X and Ydirections) at a frequency. Then the data were transferred and analyzed by multi-sensor datafusion method including Kalman filter(KF) and multi-sensor joint probabilistic dataassociation(MSJPDA) algorithm, which were studied and realized by VC++6.0. Further, thenumber, speed and type of vehicles were calculated and recognized. At the same time, fourkinds of network topologies of WSN were discussed, and the recognition ratio of effectiveinformation was analyzed separately in order to test which one had the best perfectperformance for traffic flow detection. Finally, the results showed that the algorithm for trafficflow detection had a high detection rate, and the best network topology was linear.
展开▼
