LI Minghao, ZHANG Hua, LIU Manlu, et al. Fault tolerant control method of manipulator based on deep reinforcement learning[J]. 2020, (1): 53-55+59. DOI: 10.13873/J.1000-9787(2020)01-0053-03.
Aiming at the case of sudden single joint failure of the manipulator
a fault-tolerant control method based on deep reinforcement learning is proposed. On the basis of setting up environment model and rewardpenalty mechanism
model-free reinforcement learning algorithm is used to train offline in the case of normal operation and failure operation of the manipulator. Mechanical arm model is built in Rviz and online controlled is carried out by using the above network. Experiments show that the algorithm can effectively complete the normal control and fault tolerant control of the manipulator.