多传感器交互滤波算法

 由于传感器节点感知范围有限,传感器网络内的目标跟踪过程可以被建模成为一个马尔可夫跳变系统.以此为基础根据贝叶斯理论设计接力卡尔曼滤波算法,重构新息方程,实现网络中连续的协作式跟踪.进而通过混合每次迭代状态和方差的初始值,提出了多传感器交互滤波算法.其性能优于接力卡尔曼滤波算法,却牺牲了算法的计算复杂度.最后,仿真结果验证了所提算法的有效性.

Interacting Multiple Sensor Filter for Sensor Networks

Due to the limited sensing range for sensors,moving target tracking has to be realized by relaying from one sensor to the other in sensor networks.Thus,the tracking procedure can be modeled as a Markovian chain system.By reconstructing the innovation equation,the relaying Kalman filter(RKF) algorithm is designed in the light of the Bayesian theory.On this basis,the interacting multiple sensor filter(IMSF) algorithm is proposed further by mixing the initial state and covariance at one cycle,which has a bit better tracking performance than the RKF algorithm,but at the cost of the computational complexity.Finally,simulation results show the effectiveness of the proposed algorithms.