Six-Degree-of-Freedom Robotic Arm Path Planning Algorithm Improvement And Reinforcement Learning Fusion Method

In modern industrial production,medical surgery,scientific research and other fields,the use of six-degree-of-freedom robotic arms to complete highly repetitive,high-risk and high-precision tasks has become a trend.In addition,while the development and application of robotics is advancing,the robot arm path planning problem has become a very important research direction in the field of robotics.The paper focuses on the six-degree-of-freedom robotic arm path planning problem,aiming to improve the path planning accuracy and efficiency of the robotic arm and further promote the development and application of robotics.Firstly,in the paper,the researchers proposed an improved A* algorithm by improving the cost function,heuristic function,search step size and action cost to address the problem that the A* algorithm is time-consuming and not highly accurate,and conducted several sets of comparison tests in 2-dimensional obstacle space and 3-dimensional obstacle space to analyze the experimental data to verify the effectiveness of the improvement measures.Secondly,in order to solve the node redundancy and resource waste problems of Rapidly-exploring Random Tree(RRT)algorithm,the paper proposes the Biased RRT* algorithm and combines different obstacle spaces with RRT algorithm and its derivative algorithms to conduct multiple sets of experiments.The experimental results and experimental data verify the superiority of the improved algorithm.Finally,in the paper,the researchers summarize the attempt to use the DQN algorithm to fuse the improved A* algorithm and Biased RRT* algorithm.The algorithm reduces the difficulty of algorithm implementation by transforming the continuous control problem into a discrete control problem,and creates a robot arm simulation obstacle avoidance environment in MATLAB platform,and conducts simulation experiments comparing the performance index with the above algorithms.The experimental results show that the fusion algorithm has the best performance index,the shortest time used and the highest node utilization.In summary,this thesis provides an in-depth discussion of the robotic arm path planning problem and proposes a practical and improved algorithm.These research results have certain reference value for promoting the further development and application of robotics,and can provide valuable ideas and theoretical support for research in related fields.