Traditional gait control methods are generally based on dynamics and mathematical models. Although different methods have been proposed to control biped gait, including model-based methods ( Chevallereau et al., 2013), bio-inspired methods ( Liu et al., 2020), and machine-learning-based methods ( Wang, Chaovalitwongse & Babuska, 2012), gait controlling of biped robots remains one major task for robotics. However, the gait planning of biped robots is more difficult due to the extremely complex dynamics properties. Biped robots have more lifelike movements and can walk on more complex terrains, showing great potential in nursing, rescue, and other various applications. However, wheeled robots are not capable of complex outdoor environments where the ground may slope or be uneven. Wheeled robots have the simplest moving pattern. For mobile robots, the controlling task includes both motion and moving, which increases the complexity of the design. Another family of robots, mobile robots, are widely used to perform tasks including serving, rescue, and medical treatment. Currently, various motion controlling and optimization methods have been proposed and widely used ( Sucan, Moll & Kavraki, 2012 Huda et al., 2020 Ratliff et al., 2009 Liu & Liu, 2021). For those non-moving robotics, motion planning is to control all the joints of the robot to achieve the target position. Robotic arms are widely used in factories, significantly improving productivity. Different environments are introduced including plains, slopes, uneven terrains, and walking with external force, where the robot was expected to maintain walking stability with ideal speed and little direction deviation, to validate the performance and robustness of the proposed method.Īutonomous robots have been used to perform different tasks and help to reduce workloads. In task (II), the proposed method achieved a tracking accuracy of over 95%. To our knowledge, our work achieved the best velocity performance on the platform Darwin-op. The velocity performance achieved 2× compared with the rated max velocity and more than 8× compared with other recent works. The directional accuracy improved by 87.3%. In task (I), the proposed method resulted in the walking velocity of 0.488 m/s, with a 5.8 times improvement compared with the original traditional reference controller. To validate the proposed method, the Darwin-op robot was set as the target platform and two different tasks were designed: (I) Walking as fast as possible and (II) Tracking specific velocity. Several improvements were implemented to further improve the training efficiency and performance including: random state initialization (RSI), the noise of joint angles, and a novel improvement based on symmetrization of gait. The RL environment was finely crafted for optimal performance, including the pruning of state space and action space, reward shaping, and design of episode criterion. The proposed trained agent outperformed the reference motion and existing motion-based methods. Learn and outperform the reference motion (LORM), an RL based framework for gait controlling of biped robot is proposed leveraging the prior knowledge of reference motion. However, RL algorithms are limited by the problems of convergence and training efficiency due to the complexity of the task. In addition, the high autonomy of the RL controller results in a more robust response to complex environments and terrains compared with traditional controllers. Reinforcement learning (RL) helps to overcome the complications of dynamics design and calculation. However, traditional motion controllers suffer from extremely complex dynamics properties. Legged robots are better able to adapt to different terrains compared with wheeled robots.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |