Research On Collision-free Fast Path Planning Of Super Redundant Manipulator Based On Reinforcement Learnin

Manipulators are widely used in many fields.In order to complete a certain task,the manipulator needs to complete a series of actions(configurations).The selection and design of these configurations are called the path planning of the manipulator.The hyper-redundant manipulator is more adaptable to complex environments because of its many degrees of freedom and flexibility,but it also makes path planning very difficult.At present,the path planning of hyper-redundant manipulators is often designed based on reinforcement learning.Although the use of reinforcement learning to plan paths reduces the difficulty of path planning to a certain extent,the existing methods usually first use the interaction between actions and the environment to fully understand This method has the following problems:(1)When the working environment of the manipulator changes,the model needs to be retrained,so it cannot be applied to the dynamic environment;(2)Due to the direct calculation of all actions from the initial state to the final state,the algorithm complexity is high,and one planning often takes several hours,which limits the practical application of the algorithm.In order to solve the above problems,this paper proposes a new path planning algorithm inspired by the snake crawling motion:(1)Under the condition of no collision,the trajectory configuration of the end-effector of the hyper-redundant manipulator is first planned based on reinforcement learning.(2)imitating the process of snake crawling,controlling the end-effector of the hyper-redundant manipulator to pass through the previously planned motion trajectory configuration of the end-effector,in turn,to obtain all actions from the initial configuration to the final configuration.The third chapter of this paper introduces the overall architecture and innovations of the path planning algorithm.The fourth chapter introduces the crawling motion control method of the hyper-redundant manipulator imitating the snake and analyzes the problem that the end-effector does not always coincide with the motion trajectory configuration caused by this control method.The fifth chapter introduces the reinforcement learning-based algorithm that uses self-motion obstacle avoidance to plan the trajectory configuration of the end-effector.In order to verify the effectiveness of the method,the sixth chapter of this paper conducts simulation experiments in two environments.In the pipeline environment,two kinds of hyper-redundant manipulators with 16 degrees of freedom and 24 degrees of freedom are used for experiments.The results show that the success rate of the algorithm can reach 83.3% and 96.7%,respectively,and the average time is 17.68 s and10.54 s.In the concentric pipeline environment,experiments are carried out on a 24-degree-of-freedom hyper-redundant manipulator.The results show that the path planning success rate of the algorithm is 96.1%,and the average time is 52.61 s.In summary,the experiments show that the algorithm can complete the path planning task of the hyper-redundant manipulator in a complex environment.