中国邮电高校学报(英文) ›› 2014, Vol. 21 ›› Issue (6): 78-86.doi: 10.1016/S1005-8885(14)60348-4

• Others • 上一篇    下一篇

Mobile robot SLAM method based on multi-agent particle swarm optimized particle filter

唐贤伦1,李腊梅2,蒋波杰1,3   

  1. 1. 重庆邮电大学
    2. 重庆邮电大学自动化学院(现为先进制造学院)工业工程专业
    3. 重庆邮电大学自动化学院工业工程
  • 收稿日期:2014-04-12 修回日期:2014-09-20 出版日期:2014-12-31 发布日期:2014-12-31
  • 通讯作者: 李腊梅 E-mail:sophiellm@sina.com
  • 基金资助:

    国家自然科学基金;重庆市自然科学基金

Mobile robot SLAM method based on multi-agent particle swarm optimized particle filter

  • Received:2014-04-12 Revised:2014-09-20 Online:2014-12-31 Published:2014-12-31
  • Contact: La-Mei LI E-mail:sophiellm@sina.com
  • Supported by:

    ;National Natural Science Foundation of Chongqing

摘要:  To overcome particle impoverishment, a simultaneous localization and mapping (SLAM) method based on multi-agent particle swarm optimized particle filter (MAPSOPF) was presented by introducing the idea of multi-agent to the particle swarm optimized particle filter (PSOPF) which is an algorithm for SLAM. In MAPSOPF, agents can communicate and compete with each other and learn from each other. The MAPSOPF algorithm can update the prediction of particle, adjust the proposal distribution of particles, improve localization precision and fault tolerance, and propel the particles to concentrate on the robot's true pose. Compared with standard particle filter (PF), the proposed method can achieve better SLAM precision by fewer particles. Simulations verify its effectiveness and feasibility.

关键词: PF, multi-agent system (MAS), particle swarm optimization (PSO), SLAM

Abstract:  To overcome particle impoverishment, a simultaneous localization and mapping (SLAM) method based on multi-agent particle swarm optimized particle filter (MAPSOPF) was presented by introducing the idea of multi-agent to the particle swarm optimized particle filter (PSOPF) which is an algorithm for SLAM. In MAPSOPF, agents can communicate and compete with each other and learn from each other. The MAPSOPF algorithm can update the prediction of particle, adjust the proposal distribution of particles, improve localization precision and fault tolerance, and propel the particles to concentrate on the robot's true pose. Compared with standard particle filter (PF), the proposed method can achieve better SLAM precision by fewer particles. Simulations verify its effectiveness and feasibility.

Key words: PF, multi-agent system (MAS), particle swarm optimization (PSO), SLAM

中图分类号: