基于量子PSO免疫粒子滤波的FastSLAM算法  

An Improved Fast SLAM Algorithm Based on Quantum Particle Swarm Optimization Immune Particle Filter

作　　者：潘爽[1] 刘海光[1] 李楠 聂永芳[1] PAN Shuang;LIU Hai-guang;LI Nan;NIE Yong-fang(Department of Strategic Missile and Underwater Weapon,Naval Submarine Academay,Qingdao Shandong 266199,China)

机构地区：[1]海军潜艇学院战略导弹与水中兵器系,山东青岛266199

出　　处：《计算机仿真》2018年第8期202-205,共4页Computer Simulation

基　　金：国家自然科学基金项目(61473306)

摘　　要：Fast SLAM算法中粒子滤波存在粒子退化问题,重采样可以抑制粒子退化,却又带来了多样性减弱问题。提出量子粒子群优化免疫粒子滤波的Fast SLAM算法,利用量子粒子群优化算法,减缓粒子权值的退化,再通过人工免疫算法的变异操作扩大搜索范围,从而增加粒子的多样性,避免算法陷入局部最优,提高全局搜索能力。在无人机环境下对提出的算法进行仿真验证,结果表明在SLAM的估计精度和一致性方面,提出的算法都明显优于基于重要性重采样的Fast SLAM算法。In FastSLAM algorithm, the particle degeneration always exists in particle filter. Although resampling can restrain the particle degeneration, attenuate particle diversity may be weakened. Therefore, this article presents an improved FastSLAM algorithm based on quantum particle swarm optimization immune particle filter. Firstly, the quantum particle swarm optimization algorithm was used to slow down the degeneration of particle weight. Moreover, the mutation operation of artificial immune algorithm was used to expand the search scope, so as to increase particle diversity and improve the global search ability. Thus, the local optimization was avoided. In unmanned aerial vehicle environment, the proposed algorithm was simulated. Simulation results show that the proposed method is better than the FastSLAM algorithm based on importance resampling in estimation accuracy and consistency of SLAM.

关 键 词：量子粒子群算法 人工免疫算法 粒子滤波 

分 类 号：V24[航空宇航科学与技术—飞行器设计]