无线传感器网络分布式量化卡尔曼滤波

Distributed quantized Kalman filtering for wireless sensor networks

本文针对无线传感器网络中的目标跟踪问题,研究了分布式量化卡尔曼滤波问题.由于网络中存在能量和带宽限制,传感器传输的数据必须经过量化处理.考虑一个线性离散随机动态系统,首先提出了一种动态Lloyd-Max量化器并设计了其在线更新方案,然后基于贝叶斯原理导出了递归形式的最优量化卡尔曼滤波器,同时给出了一种渐近等价的迭代算法,并进一步分析了量化卡尔曼滤波器的稳定性.最后,仿真结果验证了所设计算法的可行性与有效性.

We study the distributed quantized Kalman filtering for the target-tracking in wireless sensor networks（WSNs）.Because of the constraints on power and bandwidth in WSNs,sensor data have to be quantized before transmis-sion.A linear discrete-time stochastic dynamic system is employed for this purpose.First,a dynamic Lloyd-Max quantizer is adopted and the corresponding online update scheme is designed.Then,the optimal recursive quantized Kalman filter is derived based on the Bayesian principles,and an asymptotically equivalent iterative algorithm is developed.The sta-bility of the quantized Kalman filter is analyzed.Simulation results show the feasibility and effectiveness of the designed algorithms.