摘要
针对室内陪护机器人粒子滤波定位方法,研究了四种粒子滤波重采样算法:多项式重采样算法、残差重采样算法、分层重采样算法和系统重采样算法,并分别对其进行仿真比较。实验证明残差重采样算法粒子收敛速度和粒子匮乏程度取折衷,性能优于其它三种重采样算法,在此基础上利用仿真实验结果在HHR-0303服务机器人上进行了实验。实验证明采用残差重采样算法的粒子滤波算法,利用声纳配合里程计定位的方案能达到定位目的。
Four of the particle filter resampling algorithms in indoor service robot localization is described in this paper, they are Multinomial Resampling,Residual Reampling,Stratified Resampling and Systematic Resampling. The simulation and comparison is also presented. It is proved the performance of the Residual Resampling is better than other three algorithms on the particle convergence speed and the pinch degree. The experiment has been done on the HHR - 0303 service robot. The experiment proved that the localization plan which introduces the Residual Resarnpling 81gorithm localizing with odometer and sonar can achieve the localizing aim.
出处
《计算机技术与发展》
2010年第4期54-57,共4页
Computer Technology and Development
基金
国家高技术研究发展计划项目(2006AA040202)
关键词
陪护机器人
定位
粒子滤波
重采样算法
service robot
localization
particle filter
resampling algorithm