基于传感网的地面运动车辆目标容错跟踪

Fault tolerant-tracking of moving vehicles using sensor networks

提出了一种新的基于声震传感网的机动目标跟踪算法,即在Rao-Blackwellized蒙特卡洛数据关联(RBMCDA)算法基础上,引入代价函数,根据代价函数的可信度和误差偏离度实时在线更新测量噪声模型参数.仿真结果表明:相比于RBMCDA算法,该算法不依赖于观测噪声的精确建模,在节点频繁切换情况下仍具有很好的自适应性;相比于代价参考粒子滤波算法,在错误测量概率达10%情况下,算法仍能精确跟踪,具有很好的收敛性和容错能力;在测量噪声方差由0.001变到0.1过程中,算法能动态调整模型参数,具有较好的鲁棒性.

A tracking algorithm for moving vehicles based on acoustic and seismic sensor networks was proposed. This novel algorithm was based on Rao-Blackwellized Monte Carlo data association （RBM- CDA） algorithm, introduced cost function, and used the credibility of cost function and the error deviation to update the measurement noise model parameter online. Simulation results show that the new algorithm does not rely on the accurate model of measurement noise, and has better adaptability in case of frequent nodes switching compared with the RBMCDA algorithm; moreover, the algorithm can track accurately under 10~~ error measurement, so it has better convergence and better fault-toler- ance compared with the cost-reference particle filter algorithm; and when the measurement noise vari- ance changes from 0. 001 to 0. 1, this algorithm can dynamically adjust the model parameters with good robustness.