多障碍环境中基于增强式学习的势场优化和机器人路径规划Potential Field Optimization and Robot Path Planning in Multi-Obstacle Environment Based on Reinforcement Learning
庄晓东,孟庆春,王汉萍,殷波
摘要(Abstract)：
该文把增强式学习方法应用于多障碍环境中机器人路径规划 ,并将增强式学习和路径规划相结合 ,通过工作空间势场的自适应优化学习 ,实现机器人的全局路径规划 ,即得到从任何初始位置开始的最优路径。与传统的人工势场方法相比 ,该方法避免了势场中局部极小点所引起的陷阱区域 ,并且所得到的路径具有最优特性。计算机仿真实验结果表明 ,这种学习方法能有效的解决多障碍环境中的机器人路径规划问题
关键词(KeyWords)： 增强式学习;移动机器人;多障碍环境;人工势场;路径规划
基金项目(Foundation): 高等学校重点实验室访问学者基金;; 青岛市科委课题资助
作者(Author): 庄晓东,孟庆春,王汉萍,殷波
DOI: 10.16441/j.cnki.hdxb.2001.06.022
参考文献(References)：
- 1 MengQingchun,ZhouChangjiu,SngHongliang,etal.Study on real-timeworkspacemodelization formulti-robotsystem.[C]IEEE international conference on intelligentvehicles.Stuttgart:[s.n.],1998.569～573
- 2 RamA,ArkinR,GaryB,et al.Using genetic algorithms to learn reactive control parameters for autonomousrobotic navigation.[J]Adaptive behavior,1994,2(3):277～304
- 3 庄晓东,孟庆春,殷 波,等.动态环境中基于模糊概念的机器人路径搜索方法.[J]机器人,2001,23(5):397～399
- 4 AndersonC W.Strategy learning with multi-layerconnectionistrepresentation.[C]Proceedings of theForthIn-ternationalWorkshop onMachineLearning.Irvine:[s.n.],1987
- 5 KaelblingL P,LittmanM L,MooreA W.Reinforcementlearning:a survey.[J]Journal of artificial intelligenceresearch,1996,4:237～285
- 6 WilliamsR J.Reinforcement-learning connectionistsystems.[R]TechnicalReportNU-CCS-87-3,Boston:Col-lege ofComputerScience,NortheasternUniversity,1987
- 7 BartoA G.Learning by statistical cooperation of self-interested neuron-like computing elements.[J]Human neu-robiology,1985,(4):229～256
- 8 SamuelA L.Some studies in machine learning using the game of checkers.[J]IBM Journal onResearch andDe-velopment,1959,(3):210～229
- 9 BartoA G,SuttonR S,WatkinsC JC H.Learning and sequential decision making.[M]Learning andComputa-tionalNeuroscience:Foundations ofAdaptiveNetworks.Cambridge:MIT Press,1990.539～602
文章评论(Comment)：
|
||||||||||||||||||
|
||||||||||||||||||