基于改进的鲁棒机器人蒙特卡罗定位算法  

Algorithm Based on Improved Robust Robot Monte Carlo Localization

作　　者：杨居义[1] 

机构地区：[1]绵阳职业技术学院计算机科学系,四川绵阳621000

出　　处：《微电子学与计算机》2011年第8期150-153,共4页Microelectronics & Computer

基　　金：四川高等职业教育研究中心资助项目(GZY11B17)

摘　　要：针对粒子滤波过程的粒子退化问题和提高粒子的细化能力,提出一种基于改进的鲁棒机器人蒙特卡罗定位(Improved Robust Robot Monte Carlo localization,IRR-MCL)算法.首先利用扩展卡尔曼滤波来精确设计粒子滤波器的提议分布,将当前观测信息融入顺序重要性采样过程,以改善滤波效果,减小所需粒子数;然后,给出IRR-MCL定位算法的实现细节,实验结果表明,该算法与传统的方法在定位精度和鲁棒性方面都有显著提高.This paper puts forward a kind of algorithm based on Improved Robust Robot Monte Carlo Localization（IRR-MCL）,with the view of researching the particle degeneracy phenomenon and improving the particle refining performance in the process of the particle filter.Firstly,distribute using the proposal about accurately designing a particle filter by the extended Kalman filter and blend the current observational information into the sampling process of progression importance in order to improve the filtering effect and reduce the required particles.Then present the implementation details of IRR-MCL algorithm.Experimental results show that this algorithm,compared with the traditional methods,has improved significantly in terms of localizing accuracy and robustness.

关 键 词：机器人 蒙特卡罗定位 粒子滤波 重采样 鲁棒性 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]