Energy - Optimized Wireless Sensor Network Distributed Filtering And Fusion

Wireless sensor network is a multi-hop and self-organized wireless network formed with a large amount of wireless sensor nodes. Considering the energy, data processing, storage and communication capability of the wireless sensor node are very limited, one of the most important goals in routing protocols schedule is that how to design an effective distribution fusion algorithm in order to prolong the lifetime of the network to the maximum on the condition that the given precision is assured.Several types of distributed filter algorithms are proposed from the view of structure of the network, convergence speed, and energy-efficient utilization in this paper to deal with the limited power issue of wireless sensor network nodes. The main research contents are summarized as follows:1) A new distributed cluster-based Kalman consensus filtering algorithm is proposed to save node’s energy. In order to reducing the amount of data transmission and shorten the communication distance among nodes, the network is divided into several different clusters. Furthermore, a Gossip Algorithm is embedded into distribute Kalman filtering for reducing power consumption. Then, the convergence of the algorithm is analyzed based on graph theory and matrix theory.2) Distributed Kalman consensus filtering algorithm based on site percolation mode is studied. In order to guarantee the connectivity of the network, a new energy efficient distributed Kalman filtering algorithm is proposed based on clustering and site percolation model-- a kind of random graph. Then, the convergence of this new algorithm is proved by applying to Lyapunov second method and matrix theory.3) A fast consensus algorithm of multi-agent systems with double gains regulation is also discussed. Considering the fast consensus problem of first order discrete multi-agent systems, a class of fast consensus algorithm based on local information feedback method is proposed. Based on matrix theory and the frequency-domain analysis, two sufficient conditions for fast consensus and corresponding expressions of the convergence rate are presented respectively. Finally, a simulation example is given to verify the feasibility of the method.4) The distributed Kalman consensus filtering algorithm applying for maneuvering target track is studied. This algorithm use regular hexagon grid as the clustering model to cover the global tracking filed. Through activing some suitable clusters to detect the moving target dynamically, we can obtain state estimates of this target. Furthermore, the concept "virtual cluster heads" is introduced to create the cluster head election strategy for balancing energy consumption among sensor nodes effectively.