| In recent years, the distributed estimation of wireless sensor networks(WSN) attract attention of many researchers in different areas. As one branch of distributed estimation,distributed consensus filter have broad application prospects such as target monitoring,tracking, searching and rescuing. So there is theoretical significance and researching value for consensus filer. There will be a deep study of the subject. The main research content structure is as follows:1. The energy optimization problem of sensor networks is considered. A consensus filter is proposed. In order to reduce the energy consumption, assuming that only part of sensor nodes are used for target state estimation, the consensus gain must adjust accordingly. Though changing the binary information between sensor nodes to adjust the consensus gain, and combine the improved consensus gain with Kalman filter to estimate.2. The optimization problems of consensus filter algorithm of sensor network is considered. A new consensus weighted filter is proposed. For a filtering algorithm that already exist, the principle of optimization is minimizing the sensor network estimation error covariance to adjust the consensus gain of the network. By using the optimized the consensus gain to update the status value of sensor nodes, it can improve the consensus estimation accuracy of Kalman filtering algorithm and the robustness of the entire network.3. There is existing packet loss for wireless sensor networks in the practical application.We propose a new distributed method of filter which based on consensus and packet loss compensation. The filtering algorithm solve two kinds of packet loss in sensor networks: for the failure of data-detection of nodes, the filtering introduces a consensus protocol and improving the network robustness for packet loss. To solve the problem of communication failure between the sensor nodes, the filtering proposes packet dropout compensation method to reduce the estimation error of the networks. |
PDF Full Text Request