Design of the Optimal Quantizer Based on the Fuzzy C-means Cluster
Yan Zeng · Journal of Sichuan Normal University · 2010
Distributed estimation fusion is an important research topic of multisensor systems.In practical application,the distribution of observation statistics of sensor is usually unknown.Meanwhile,the observation and estimation from individual sensor must be compressed and coded effectively because of communication bandwidth constraints.In the current study,utilizing direct sum estimation approach presented by V.Megalooikonomou et al.,the authors design an optimal quantizer,on the basis of the fuzzy C-means cluster,used for multi-sensor distributed estimation fusion system when the joint probability of observation statistics is unknown.The proposed method can significantly improve the fusion performance in comparison with that quantizer which is based on the hard C-means cluster.The computer simulations show the effectiveness of the presented method.