Application and Comparative Study of Optimized Momentum Gradient Descent Algorithm in Trajectory Planning of 6-DOF Manipulator

Due to the wide application of robotic arms in current industrial and other fields, optimal path planning for robotic arms has become an important research topic. There are a large number of path planning algorithms in the field, but most of them face problems such as easy to fall into local optimal solutions, path oscillation, principle complexity or difficulty in adjusting parameters. This paper proposes an optimized momentum gradient descent algorithm to search for paths in three-dimensional space, combining dynamic learning rate, joint angle update limitation, moving average filtering, etc., and taking path length, number of iterations, comprehensive indicators of speed and acceleration changes of each joint as quality judgment criteria. The experimental results show that the proposed method can effectively plan a smooth path with relatively stable angle, speed and speed changes.