无线层析网络中基于粒子滤波的时变多目标跟踪

研究将粒子滤波（PF）理论应用于无线层析网络中的时变多目标跟踪（MTT）.传统的基于无线层析成像（RTI）的时变多目标跟踪方法存在延迟问题，即与真实的时变目标数目相比，估计所得时变目标数目存在滞后，并且跟踪精度较低.基于PF的时变多目标跟踪方法利用维度可变的粒子来估计目标数目，实现目标跟踪.该方法不存在延迟问题，并且能提高目标的跟踪性能.研究通过在一个9.5 m×9.5 m的监测区域内进行实验来验证该算法的有效性.实验结果表明基于RTI的时变多目标跟踪方法的最佳子模式分配（OSPA）误差为0.485 m，而基于PF的时变多目标跟踪方法的OSPA误差为0.362 m，其性能比基于RTI的方法提高了25%. 

Time-Varying Multi-Target Tracking Method Based on Particle Filter in Radio Tomographic Network

Traditional method based on radio tomographic image (RTI) suffers from latency, resulting in the existence of a lag between the estimated target number and the true one. Moreover, the tracking accuracy of the traditional method should be improved. In this paper, a particle filtering (PF) theory was introduced for time-varying multi-target tracking (MTT) in radio tomographic network, utilizing the particles with variable dimensions to estimate the target number and track the targets to solve the latency problem and improves the tracking accuracy. Some experiments were conducted in a monitored region with the area of 9.5 m×9.5 m to verify the effectiveness of the PF-based method. The experimental results show that the optimal sub-pattern assignment (OSPA) error of traditional method is 0.485 m. In contrast,the OSPA error of proposed method is 0.362 m, which is improved by 25%.