Author(s): Zhang Zhi, Li Jianxun, Liu Liu, Liu Zhaolei, Han Shan, Department of Automation, Shanghai Jiao Tong University, Key Laboratory of System Control and Information Processing, Ministry of Education of China, Nanjing Research Institute of Electronics Technology, Engineering Research Center of Wideband Wireless Communication Technology, Ministry of Education, Nanjing University of Posts and Telecommunications
Pages: 1735-1746
Year: 2015 Issue:  6
Journal: Chinese Journal of Aeronautics
Keyword:  Posterior Crame′r–Rao lower bounds; Quantization; State estimation; Target tracking; Wireless sensor networks;
Abstract: Since the features of low energy consumption and limited power supply are very important for wireless sensor networks(WSNs), the problems of distributed state estimation with quantized innovations are investigated in this paper. In the first place, the assumptions of prior and posterior probability density function(PDF) with quantized innovations in the previous papers are analyzed. After that, an innovative Gaussian mixture estimator is proposed. On this basis, this paper presents a Gaussian mixture state estimation algorithm based on quantized innovations for WSNs. In order to evaluate and compare the performance of this kind of state estimation algorithms for WSNs, the posterior Crame′r–Rao lower bound(CRLB) with quantized innovations is put forward. Performance analysis and simulations show that the proposed Gaussian mixture state estimation algorithm is efficient than the others under the same number of quantization levels and the performance of these algorithms can be benchmarked by the theoretical lower bound.