Deep reinforcement learning-based rehabilitation robot trajectory planning with optimized reward functions

Deep reinforcement learning (DRL) provides a new solution for rehabilitation robot trajectory planning in the unstructured working environment, which can bring great convenience to patients. Previous researches mainly focused on optimization strategies but ignored the construction of reward functions, which leads to low efficiency. Different from traditional sparse reward function, this paper proposes two dense reward functions. First, azimuth reward function mainly provides a global guidance and reasonable constraints in the exploration. To further improve the efficiency, a process-oriented aspiration reward function is proposed, it is capable of accelerating the exploration process and avoid locally optimal solution. Experiments show that the proposed reward functions are able to accelerate the convergence rate by 38.4% on average with the mainstream DRL methods. The mean of convergence also increases by 9.5%, and the percentage of standard deviation decreases by 21.2%–23.3%. Results show that the proposed reward functions can significantly improve learning efficiency of DRL methods, and then provide practical possibility for automatic trajectory planning of rehabilitation robot.

Recent Advances in Robot Trajectory Planning in a Dynamic Environment

Background: Trajectory planning is important to research in robotics. As the application environment changes rapidly, robot trajectory planning in a static environment can no longer meet actual needs. Therefore, a lot of research has turned to robot trajectory planning in a dynamic environment. Objective: This paper aims at providing references for researchers from related fields by reviewing recent advances in robot trajectory planning in a dynamic environment. Methods: This paper reviews the latest patents and current representative articles related to robot trajectory planning in a dynamic environment and introduces some key methods of references from the aspects of algorithm, innovation and principle. Results: In this paper, we classified the researches related to robot trajectory planning in a dynamic environment in the last 10 years, introduced and analyzed the advantages of different algorithms in these patents and articles, and the future developments and potential problems in this field are discussed. Conclusion: Trajectory planning in a dynamic environment can help robots to accomplish tasks in a complex environment, improving robots’ intelligence, work efficiency and adaptability to the environment. Current research focuses on dynamic obstacle avoidance, parameter optimization, real-time planning, and efficient work, which can be used to solve robot trajectory planning in a dynamic environment. In terms of the combination of multiple algorithms, multi-sensor information fusion, the combination of local planning and global planning, and multi-robot and multi-task collaboration, more improvements and innovations are needed. It should create more patents on robot trajectory planning in a dynamic environment.