LUO Ying, TAN Guan-zheng. Research on global localization of mobile robot based on SSPF algorithm[J]. 2008, (5): 31-34. DOI: 10.13873/j.1000-97872008.05.018.
基于SSPF算法的移动机器人全局定位研究
摘要
传统粒子滤波器(PF)算法在移动机器人全局定位时会出现粒子退化和粒子耗尽问题
导致定位失败。提出了一种改进的PF算法即智能采样PF(SSPF)算法
该方法采用Unscented卡尔曼滤波器(UKF)算法和基于相关熵的自适应重采样算法相结合的方法
较好解决了传统PF算法所引起的粒子退化问题。由于重采样的一种极端情况会带来粒子耗尽问题
采用MH(metropolis hastings)算法
较好地解决了粒子耗尽问题。与以往方法相比
该方法可以实现准确性较高的移动机器人全局定位。
Abstract
Traditional particle filter(PF) algorithm can lead to particle degeneracy and particle depletion problems in global localization of mobile robot to cause localizing defeat.A improved PF named smart sampling particle filter(SSPF) algorithm is proposed by combining Unscented Kalman filter(UKF) algorithm and relative entropy for adaptive PF algorithm
which solves the particle degeneracy of traditional PF algorithm.Since one of the extreme conditions of resampling can cause particle depletion
this problem is solved by using metropolis hastings(MH) approach.Comparing with previous methods
the algorithm can implement global localization of mobile robot with high precise.
关键词
Keywords
references
粒子滤波算法综述 [J]. 胡士强,敬忠良. 控制与决策 . 2005(04)
Robust Monte Carlo localization for mobile robots [J] . Sebastian Thrun,Dieter Fox,Wolfram Burgard,Frank Dellaert. Artificial Intelligence . 2001 (1)