Abstract
AbstractA policy gradient reinforcement learning method for backward motion control of tractor-trailer mobile robot is proposed in this paper. The kinematic model is built firstly. Then the control action of mobile robot is trained by reinforcement learning. Policy gradient (PG) learning algorithm is designed to train the backward control action of the TTMR system. The PG based neural network structure is constructed and the parameters of neural network are updated through the PG method. The mobile robot learns a strategy by finite iterations finally. The experimental results show that the mobile robot can reverse the trailer stably in large enough iterations. The designed learning method based on PG algorithm is effective in the motion control of mobile robot with trailer.KeywordsReinforcement learningArtificial intelligencePolicy gradientNeural networkTractor-trailer robotBackward motion
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
