Distributed Fusion Method under Imperfect Channel in Wireless Sensor Networks

Wireless sensor networks have been widely developed in modern surveillance systems since its complementary and overlapping coverage on targets. In order to track and estimate the path of target, the sensor data should to be fused. However, the information fusion performance is influenced largely due to the presence of disruptions and delays in communication channels between sensor nodes which are not sufficiently solved until now. A stochastic estimation function is proposed in this paper to predict the fusion performance by adopting the effects of network connectivity and message delay into the fusion equations. Compared with the traditional deterministic estimation function, the final equations produced by the method consist of some prior parameters which can be determined according to the specific communication environment where the sensor networks are deployed. The simulation results have shown that the stochastic method can reduce the fusion time markedly in addition to provide good estimation capabilities.