Robot Trajectory Planning Optimization Algorithm Based on Improved TD3 Algorithm

Abstract

In order to solve the problem that the existing manipulator is easy to fall into the local optimal state when carrying out path planning, which leads to low efficiency, this paper uses a combination of simulated annealing algorithm and deep reinforcement learning algorithm to solve the path planning problem. The algorithm inserts the solution formed by the simulated annealing algorithm exploration process into the population and combines it with the double delay determination strategy gradient. The system selects whether to accept the new solution based on the Metropolis criterion, and has a certain probability to accept the solution with poor results. A 30 × 30 maze environment was built under python, and the training route was carried out. The experimental results show that when the learning rate is 0.8 and the discount factor is 0.95, the SA-TD3 algorithm has faster convergence speed and higher efficiency than the DDPG and TD3 algorithms.