Decentralized collaborative self-organization algorithm based on information content

Collaborative self-organization of sensor nodes in wireless sensor networks involves sensor management and state estimation, which include the selection of sensor nodes, the configuration of sensors, and the estimation of the states of the inspected system. Thus, this collaborative self-organization performs the joint optimization of decision and estimation. We propose an adaptive dynamic collaborative self-organization algorithm, in which the sensors are selected based on the composite index of the measured information and the residual energy of the sensor node. Given the desired accuracy by the end user, the optimal set of sensors involved in the sensing task is chosen adaptively and instantly. Then, the measured information from selected sensors is fused under the frame of information filter. Compared with the method of information-driven sensor querying (IDSQ), this technique is more advantageous in the adjustable accuracy, the robustness and the network lifetime. When this algorithm is applied to the target tracking, the simulation results validate the superiority of this algorithm to IDSQ in tracking accuracy, the percentage of missing tracking and the lifetime of the network.