Abstract: Hierarchical wireless sensor networks (WSNs) consist of backbone nodes and normal sensor nodes. Because of limited energy resource, tracking targets in WSN should consider both the tracking accuracy and energy consumption. A multi-step optimized measurement approach based on dynamic clustering is proposed. The backbone node implements particle filter to predict target locations, and selects normal sensor nodes around the target to be measurement sensors based on DELAUNAY triangulation. It then judges whether the cluster head role needs to be transferred. When the measurement is done, the backbone node selects the sensor with the most energy-efficiency, which is computed by using entropy, to localize the target step by step. The experiments verify that the proposed approach can well satisfy the target tracking accuracy, and reduce energy consumption efficiently.

Key words: Dynamic clustering, Hierarchical, Multi-step optimization, Wireless sensor networks