|
|
|
| Improved FastSLAM algorithm based on importance weight smoothing |
| CHEN Jia-qian, HE Yan, JIANG Jing-ping |
| College of Electrical Engineering, Zhejiang University, Hangzhou Zhejiang 310027, China |
|
|
|
Abstract Frequent resampling during the process of the FastSLAM algorithm leads to quick sample impoverishment that will subsequently cause the loss of landmark estimate diversity and affect the final estimate result. A novel improved FastSLAM algorithm based on importance weight smoothing was proposed to overcome the problem. Not only the current motion and observation information but also all the past importance weights in a sliding window influenced the current importance weights that were calculated by smoothing. So by reducing the over fluctuation of the importance weights induced by noise and normalization, the proposed method decreased the resampling times and improved the estimate effect. Monte Carlo simulations in two different environments indicate that with an appropriate sliding window the proposed method can effectively reduce the resampling frequency and achieve better estimation precision．
|
|
Published: 21 September 2010
|
|
|
基于权值平滑的改良FastSLAM算法
针对FastSLAM算法中频繁重采样会导致粒子快速坍塌,从而破化路标估计的多样性并最终影响估计结果的问题,提出一种基于权值平滑的改良算法.该方法采用平滑方式计算粒子权值,不仅考虑机器人当前的运动和观测结果,并且综合一定长度滑动窗口内的历史权值信息,可抑制由噪声和归一化等因素引起的权值过度波动,以及由此引发的频繁重采样和估计性能降低.蒙特卡罗仿真结果表明,选取合适的滑动窗口大小,改良算法能有效减少重采样次数,保持粒子多样性,显著提高估计精度.
|
|
| ［1］ DURRANTWHYTE H F, RYE D, NEBOT E. Localization of autonomous guided vehicles ［C］∥ Proceedings of the 8th International symposium on Robotics Research, New York: SpringerVerlag, 1995: 613625．
［2］ SMITH R C, CHEESEMAN P. On the representation and estimation of spatial uncertainty ［J］. The International Journal of Robotics Research, 1986, 5(4): 5668．
［3］ SMITH R, SELF M, CHEESEMAN P. Estimating uncertain spatial relationships in robotics ［M］. Autonomous Robot Vehicles, New York: SpringerVerlag, 1990: 167193．
［4］ MURPHY K. Bayesian map learning in dynamic environments ［C］∥ Advances in Neural Information processing Systems. Denver: ［s.n.］, 1999: 10511021．
［5］ MONTEMERLO M. FastSLAM: a factored solution to the simultaneous localization and mapping problem with unknown data association［D］. Pittsburgh: Carnegie Mellon University, 2003．
［6］ MONTEMERLO M, THRUN S, KOLLER D, et al. FastSLAM 20: an improved particle filtering algorithm for simultaneous localization and mapping that provably converges ［C］∥ International Joint Conference on Artificial Intelligence. Acapulco: ［s. n. ］, 2003: 11511156．
［7］ BAILEY T, NIETO J, NEBOT E. Consistency of the FastSLAM algorithm ［C］∥ Proceedings of the IEEE International Conference on Robotics and Automation. Orlando: IEEE, 2006: 424429．
［8］ CRISAN D, DOUCET A. A survey of convergence results on particle filtering methods for practitioners ［J］. IEEE Transactions on Signal Processing, 2002, 50(3): 736746．
［9］ DOUCET A. On sequential simulationbased methods for Bayesian filtering ［R］. Cambridge Cambridge University, 1998．
［10］ GRISETTI G, STACHNISS C, BURGARD W. Improving gridbased SLAM with RaoBlackwellized particle filters by adaptive proposals and selective resampling ［C］∥ Proceedings of the IEEE International Conference on Robotics and Automation. Barcelona: IEEE, 2005: 24432448．
［11］ LIU J S. Metropolized independent sampling with comparisons to rejection sampling and importance sampling ［J］. Statistics and Computing, 1996, 6:113119．
［12］ CUGLIARI M, MARTINELLI F. A FastSLAM algorithm based on the unscented filtering with adaptive selective resampling ［C］∥ Springer Tracts in Advanced Robotics. New York: SpringerVerlag, 2008, 42: 359368．
［13］ 沈庭芝, 王卫江, 闫雪梅. 数字图像处理及模式识别 ［M］. 2版.北京: 北京理工大学出版社, 2007:5053． |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
