基于粒子滤波的移动物体定位和追踪算法

提出一种基于粒子滤波的目标定位算法PFTL(particle filter based target localization)以及一种基于网络覆盖问题的节点组织策略SAC(sampling aware tracking cluster formation).PFTL 的基本思想是,采用一系列带权粒子(weighted particles)来预测移动物体位置的后验分布空间,每个新时刻根据传感器的测量数据来权衡和定位目标.PFTL 通过引入误差容忍(error tolerant)的方式来存储和发送目标位置数据,使汇聚点关于物体位置信息的数据误差在一个可控的范围内,进而极大地减少网络通信负荷.SAC基于传感器采样离散化的特点来制订数据融合策略,并以最大化覆盖物体运动轨的方式动态地选取节点和进行节点簇的有效组织.模拟实验结果表明,与现有的几种定位算法和追踪协议相比,结合PFTL 算法和SAC 策略能够以较小的代价取得更好的定位效果和网络负载均衡,进而延长网络寿命.

Moving Target Localization and Tracking Algorithms: A Particle Filter Based Method

This paper proposes a particle filter based target localization (PFTL) algorithm, and a sampling aware tracking cluster formation (SAC) scheme for organizing the sensor nodes, which maximizes the coverage area of target's trajectory in each cluster. The key idea of PFTL is to represent the possible locations of mobile target with a number of weighted particles, and to estimate the particles for computing the position of the object when the range measurements are available at next sampling time step. The motivation behind PFTL is that if the sink is willing to tolerate a small error, regarding the position of the target, the in-network communication can be greatly decreased, as well as the consumed energy, which is the most precious resource in wireless sensor networks. To balance the computation and communication overhead of network, this study designed a node scheduling scheme SAC, which dynamically clusters the sensors aimed at minimizing the times of tracking data hand-offs, so as to save the energy expenditure. Extensive simulations are conducted to verify the proposed methodologies, and the results reveal that PFTL and SAC not only reduce the localization error, but also efficiently extend the network lifetime.