无线传感器网络环境下基于粒子滤波的移动机器人SLAM算法

定位问题是移动机器人研究领域中最基本的问题，在Bayes的框架下研究了机器人与无线传感器网络(WSN)组成系统中的同时建图与定位问题(SLAM).针对该系统中只存在距离测量信息可用的情况提出了一种基于粒子滤波的SLAM算法.该方法将机器人状态和节点位置估计设置为一组全局估计粒子，通过对粒子及其权重的更新来计算整个系统的状态.算法将WSN节点的位置估计在机器人的路径上分解为相互独立的估计，从而将全局粒子的计算转化为使用一个机器人状态滤波器和对应于每个机器人粒子的节点位置滤波器进行计算.针对观测信息低维的特点，设计了处理低维观测信息的方法，使得观测信息可以在滤波阶段得到合理利用.并且详细介绍了提出的SLAM算法原理和计算过程，并通过仿真实验证明了算法的有效性和实用性.

Simultaneous localization and mapping of a mobile robot in wireless sensor networks based on particle filtering

Localization is one of the most fundamental problems in mobile robots. A method for simultaneous localization and mapping (SLAM) in robot and WSN systems using range-only measurements was presented in a Bayes framework and a particle filtering method was designed for the problem. The estimations of the robot’s path and WSN node position were set to be clusters of particles which were called status particles. The status particles were used to estimate the whole state posterior by its position and weight. The algorithm assumed position of WSN nodes which were conditioned independently along the robot’s path, so the system posterior could be computed separately. A particle filter called a robot filter was used to compute the robot’s posterior and a separate copy of each node filter corresponding to each robot particle. Due to the low dimension of range measurement, methods were made for utilizing this information. The experiment proved the efficiency and practicality of the algorithm.