Abstract:

An agent-based D-S data fusion algorithm with evidence weights in wireless sensor networks is proposed.Evidence weights are adopted to modify the evidences before fusion in order to reduce the impact of data conflicts on fusion result.Three-level D-S combination rules are used to make fusion decision, the fusion detection probabilities of single node in time domain are calculated in the node-level fusion,and the detection probabilities of intra-cluster nodes in space domain are calculated to obtain the local decision in the intra-cluster level fusion.The fusion detection probabilities among clusters are calculated to obtain the final global decision in the inter-cluster level fusion.The stimulation results show that the proposed algorithm can obtain an accurate fusion result with lower energy consumption.

Key words: wireless sensor network;mobile agent;data fusion;D-S evidence theory;evidence weight