基于改进的粒子群优化的FastSLAM方法  

An approach to FastSLAM based on improved particle swarm optimization algorithm

作　　者：刘利枚[1,2] 蔡自兴[1] 

机构地区：[1]中南大学信息科学与工程学院智能系统与智能软件研究所,长沙410083 [2]湖南商学院计算机与电子工程学院,长沙410205

出　　处：《高技术通讯》2011年第4期422-427,共6页Chinese High Technology Letters

基　　金：国家自然科学基金（90820302,60805027）和国家博士点基金（200805330005）资助项目.

摘　　要：提出了一种基于改进的粒子群优化（IPSO）的快速同时定位和地图创建（FastSLAM）方法——IPSO FastSLAM算法。该算法在粒子预估过程中引入观测信息，调整了粒子的提议分布，增强了位置预测的准确性。改进的粒子群优化采用两步优化策略，即首先通过种群速度自适应调整惯性权重，有效地克服了粒子退化问题，改善了算法的实时性，然后针对粒子耗尽问题，在粒子群优化算法中引入遗传算法的变异运算对其进行改进，扩大解空间的范围，从而保持了种群的多样性。仿真和实时数据实验验证了该方法正确、可行。The IPSOFastSLAM algorithm, an approach to a factored solution to fast simultaneous localization and mapping （FastSLAM） based on an improved particle swarm optimization （IPSO） algorithm, is presented to improve the location of a moving robot. The algorithm incorporates the newest observation information into the prediction of particles, adjucts the proposal distribution of the particles, and the accuracy of prediction of a robot＇ s position is enhanced. The improved PSO particle swarm optimization is solved by a sequential two-step method. Firstly, the average absolute value of velocity of all particles is defined to change the inertia weight adaptively, so the degeneration of particles is overcome effectively and the real-time performance of the algorithm is improved. Then, focusing on the depletion of the particle, the mutation operation based on the genetic algorithm is adopted in the particle swarm optimization, so that the overall searching ability is enhanced, thus keeping the population diversity. The experimental results prove that the improved method is correct and feasible.

关 键 词：粒子群优化(PSO) 快速同时定位和地图创建(FastSLAM) 惯性权重 遗传算法 提议分布 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]