论文部分内容阅读
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 new maneuvering target tracking algorithm based on acoustic sensor network is proposed. Based on the Rao-Blackwellized Monte Carlo Data Association (RBMCDA) algorithm, a cost function is introduced. According to the reliability and error of the cost function The parameters of the measurement noise model are updated online in real time.The simulation results show that the proposed algorithm does not depend on accurate modeling of observed noise compared to the RBMCDA algorithm and still has good adaptability when the node frequently switches over. Particle filter algorithm, the error detection probability of 10%, the algorithm can still be accurately tracked, with good convergence and fault tolerance; in the measurement of noise variance from 0.001 to 0.1 process, the algorithm can dynamically adjust the model parameters, with Better robustness.