SelfdrivingCarNavigationwithObstacleAvoidanceUsingReinforcementLearning

SelfdrivingCarNavigationwithObstacleAvoidanceUsingReinforcementLearning.pdf
Self-drivingcarnavigationwithobstacleavoidanceproblemisahottopicinacademicworld.Thegoalofthisproblemistodesignanautonomouscarwithlearningabilitytofindacollision-freepathinanunknownenvironment.Generalsolutionstothisproblemcouldbedividedintotwopieces,navigationenvironmentdetectionandcontrolstrategydesign.Inthispaper,weusedreinforcementlearning(RL)toapproachcontrolstrategydesignandbuilttwodifferentnavigationbenchmarkswithdifferentobstaclessituationstoverifythecontrolstrategy.ThesimulationresultsshowedthattheRLprovidesaneffectivesolutiontoself-drivingcarnavigationwithobstaclesavoidanceproblem.
作者：FangXiao ChenXiaohua GaoHongbo YangMing
作者单位：CheryScientificResearchInstitute;ShanghaiJiaoTongUniversityCheryScientificResearchInstituteCheryScientificResearchInstitute;NanjingUniversityofScience&TechnologyShanghaiJiaoTongUniversity
母体文献：2015中国汽车工程学会年会论文集
会议名称：2015中国汽车工程学会年会  
会议时间：2015年10月1日
会议地点：上海
主办单位：中国汽车工程学会
语种：chi
分类号：
关键词：reinforcementlearning  carnavigation  obstacleavoidance  adaptivedynamicprogramming  reinforcementsignal
在线出版日期：2018年1月31日
基金项目：
相似文献
相关博文

SelfdrivingCarNavigationwithObstacleAvoidanceUsingReinforcementLearning.pdf

文件大小:

962.71 KB

高速下载