基于ICP算法和粒子滤波的未知环境地图创建

为了实现移动机器人仅依靠激光测距仪和里程计实时地创建精确的栅格地图, 本文提出了一种结合最近点迭代(Iterative closest point, ICP)算法和Rao-Blackwellized粒子滤波的同时定位与地图创建方法. 该方法利用ICP算法对相邻两次激光扫描数据进行配准, 并将配准结果代替误差较大的里程计读数, 以改善基于里程计读数的建议分布函数; 同时通过采用改进的抽样策略, 提高了粒子滤波过程中的抽样效率, 降低创建地图所需的粒子数. 仿真结果表明了该方法的有效性.

A SLAM Approach by Combining ICP Algorithm and Particle Filter

For building the consistent grid-based map of mobile robot only with laser range finder and odometer, this paper presents a novel algorithm that combines iterative closest point (ICP) algorithm with Rao-Blackwellized particle filter. It employs ICP algorithm to register one range scan to a previous scan so as to compute the relative robot position, then uses the result to replace the odometer reading and improve the proposal distribution. With an improved resampling method, the number of samples required is seriously reduced. Simulations on the real robot data sets illustrate the superior performance of our approach.