张晓路, 李斌, 常健, 唐敬阁. 水下滑翔蛇形机器人滑翔控制的强化学习方法[J]. 机器人, 2019, 41(3): 334-342. DOI: 10.13973/j.cnki.robot.180398
引用本文: 张晓路, 李斌, 常健, 唐敬阁. 水下滑翔蛇形机器人滑翔控制的强化学习方法[J]. 机器人, 2019, 41(3): 334-342. DOI: 10.13973/j.cnki.robot.180398
ZHANG Xiaolu, LI Bin, CHANG Jian, TANG Jingge. A Reinforcement Learning Method for Gliding Control of Underwater Gliding Snake-like Robot[J]. ROBOT, 2019, 41(3): 334-342. DOI: 10.13973/j.cnki.robot.180398
Citation: ZHANG Xiaolu, LI Bin, CHANG Jian, TANG Jingge. A Reinforcement Learning Method for Gliding Control of Underwater Gliding Snake-like Robot[J]. ROBOT, 2019, 41(3): 334-342. DOI: 10.13973/j.cnki.robot.180398

水下滑翔蛇形机器人滑翔控制的强化学习方法

A Reinforcement Learning Method for Gliding Control of Underwater Gliding Snake-like Robot

  • 摘要: 研究了一种强化学习算法,用于水下滑翔蛇形机器人的滑翔运动控制.针对水动力环境难以建模的问题,使用强化学习方法使水下滑翔蛇形机器人自适应复杂的水环境,并自动学习仅通过调节浮力来控制滑翔运动.对此,提出了循环神经网络蒙特卡洛策略梯度算法,改善了由于机器人的状态难以完全观测而导致的算法难以训练的问题,并将水下滑翔蛇形机器人的基本滑翔动作控制问题近似为马尔可夫决策过程,从而得到有效的滑翔控制策略.通过仿真和实验证明了所提出方法的有效性.

     

    Abstract: A reinforcement learning algorithm for gliding control of underwater gliding snake-like robot is studied. To solve the problem that the hydrodynamic environment is hard to be modeled, a reinforcement learning method is adopted so that the underwater gliding snake-like robot can adapt to the complex water environment and automatically learn the gliding actions only by adjusting buoyancy. A Monte Carlo policy gradient algorithm using recurrent neural network is proposed to solve the problem that the algorithm is difficult to train because the robot state can't be fully observed. The gliding action control of the underwater gliding snake-like robot is approximated as Markov decision processes (MDPs), so as to obtain an effective gliding control policy. Simulation and experiment results show the effectiveness of the proposed method.

     

/

返回文章
返回