|Table of Contents|

Mobile Robot Path Planning Based on State Sensitivity


Research Field:
Publishing date:


Mobile Robot Path Planning Based on State Sensitivity
ZHAO YunCHEN Qing-weiHU Wei-li
School of Automation,NUST,Nanjing 210094,China
mobile robots path planning reinforcement learning state sensitivity
Aiming at the mobile robot system in an unknown environment,the path planning problem to avoid both static and dynamic barriers and to reach a target quickly is investigated here.A new state sensitivity is defined to measure the relative degree between the state and the objective.It guides a robot to explore the environment with right direction and strength automatically.A reinforcement learning algorithm is adopted to learn the best action policy of a robot.By introducing state sensitivity,the speed and performance of learning algorithm are improved.Simulation results from a path planning task with an unknown environment and dynamic barriers verify the efficiency of the proposed algorithm.


[1] Mansouri M,Shoorehdeli M A,Teshnehlab M. Integer GA for mobile robot path planning with using another GA as repairing function[A]. International Conference on Automation and Logistic [C]. Qingdao,China [s. n. ], 2008: 135-140.
[2] Er M J,Zhou Y. A novel framework for automatic generation of fuzzy neural networks[J]. Neurocomputing, 2008( 71) : 584-591.
[3] CharifaS,Bikdash M. Adaptive boundary-following algorithm guided by artificial potential field for robot navigation [A]. Robotic Intelligence in Informationally Structured Space ( RIISS’09) [C]. Nashville,USA: IEEE Workshop on Robotic Intelligence in Informationally Structured Space, 2009: 38-45.
[4] 蔡文彬,朱庆保. 未知环境下基于快速搜索随机树的机器人路径滚动规划[J]. 南京师范大学学报( 工程技术版) , 2009,9 ( 2) : 79-83.
[5] Moon I H,Miura J,Shirai Y. On-line viewpoint and motion planning for efficient visual navigation under uncertainty[J]. Robotics and Autonomous Systems, 1999: 2943-2948.
[6] Jan P,Stefan S. Reinforcement learning of motor skills with policy gradients[J]. Robotics and Neuroscience, 2008, 21( 4) : 682-697.
[7] Karumanchi S,Allen T,Bailey T. Non-parametric learning to aid path planning over slopes[J]. The International Journal of Robotics Research, 2010, 29( 8) : 997-1018.
[8] 沈晶,顾国昌,刘海波. 未知动态环境中基于分层强化学习的移动机器人路径规划[J]. 机器人,2006, 28( 5) : 544-552.
[9] Whiteson S,Stone P. On-line evolutionary computation for reinforcement learning in stochastic domains[A]. Proceedings of the 8th Annual Conference on Genetic and Evolutionary Computation( GECCO 06) [C]. New York,USA: [s. n. ], 2006: 1577-1584.
[10] Kaelbling L P. Associative reinforcement learning: A generate and test algorithm[J]. Machine Learning, 1994, 15( 3) : 299-319.
[11] Singh S, Jaakkola T,Littman M, et al. Convergence results for single-step on-policy reinforcement-learning algorithms [J]. Machine Learning, 2000, 38( 3) : 287-308.


Last Update: 2012-10-12