Distributed Target Localization in Wireless Sensor Networks using Diffusion Adaptation

Localization is an important issue for wireless sensor networks. Target localization has attracted many researchers who work on location based services such as navigation, public transportation and so on. Localization algorithms may be performed in a centralized or distributed manner. In this paper we apply diffusion strategy to the Gauss Newton method and introduce a new distributed diffusion based target localization algorithm for wireless sensor networks. In our proposed method, each node knows its own location and estimates the location of target using received signal strength. Then, all nodes cooperate with their neighbors and share their measurements to improve the accuracy of their decisions. In our proposed diffusion based algorithm, each node can localize target individually using its own and neighbor’s measurements, therefore, the power consumption decreases. Simulation results confirm that our proposed method improves the accuracy of target localization compared with alternative distributed consensus based target localization algorithms. Our proposed algorithm is also shown that is robust against network topology and is insensitive to uncertainty of sensor nodes’ location.