卡尔曼一致性滤波器的丢包性能分析及能量优化

针对无线传感器网络(wireless sensor networks,WSNs)在实际应用中不可避免的数据包丢失现象,本文研究了分布式卡尔曼一致性滤波算法(distributed Kalman consensus filtering algorithm,DKF)在两类丢包情况下的稳定性和滤波性能问题,通过矩阵论理论分析得出了估计误差协方差收敛所能容忍的极限丢包率.然后,考虑到传感器节点能量有限,基于逾渗模型构建了一种能量可调的改进型分布式一致性卡尔曼滤波器,该滤波器充分利用无线传感器节点冗余布置的特点,以较小的滤波精度下降为代价,获取网络寿命的大幅度提高,实现了该分布式滤波器在滤波精度与能量消耗两个关键指标的有效权衡.最后利用仿真实例验证了所提出算法的有效性.

Packet-dropout performance and energy optimization of the distributed Kalman consensus filter

Consider packet loss happens inevitably during data exchange in the practical applications of wireless sensornetworks (WSNs), the stability and filtering performance of the distributed Kalman consensus filter (DKF) are discussedunder two classes of packet dropout in this paper, including observation packet dropout and communication one. Thoughthe matrix theoretic analysis, the upper limit of packet loss rate is derived for guarantee the convergence of estimation errorcovariance. Furthermore, consider sensor node has limited energy, an improved distributed Kalman consensus filter (IDKF)with adjustable energy consumption is proposed based on the percolation model. By full use of the characteristics of sensornode redundancy layout, the algorithm obtains longer network lifetime at the cost of less reducing filtering accuracy. So thefilter achieves the effective trade-off between the two key indicators, including filtering accuracy and energy consumption.Finally, a simulation example is given to verify the effectiveness of the proposed two algorithms.