基于模糊C-均值聚类的最优量化器设计

Design of the Optimal Quantizer Based on the Fuzzy C-means Cluster

分布式估计融合是多传感器系统研究的一项重要内容.在实际应用中,各传感器观测的统计信息往往未知,同时,由于受到通讯带宽的限制,又需要对传感器的观测和估计进行有效的压缩编码.针对观测统计量的联合概率分布未知的多传感器分布式估计融合系统,利用V.Megalooikonomou等提出的直和估计思想,基于模糊C-均值聚类方法设计一种分站最优量化器,显著地改进了基于硬C-均值聚类方法设计量化器的融合效果.计算机模拟表明了此方法的有效性.

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.