【摘要】 文章研究了强化学习在小车二阶倒立摆平衡控制中的应用。首先对小车二阶倒立摆进行建模和动力学分析；其次针对小车二阶倒立摆的平衡控制设计了基于DQN(Deep Q-Network)算法的智能体。为使控制效果更直观，基于Matlab/Simulink中的SimscapeMultibody模块搭建了二阶倒立摆物理仿真模型。仿真结果表明，基于DQN算法的智能体可以较好地控制小车二阶倒立摆的平衡，由此证明了DQN算法在二阶倒立摆稳定控制中的有效性。更多还原

【Abstract】 This paper studies the application of reinforcement learning in the balance control of a second-order inverted pendulum. Firstly, modeling and dynamic analysis of the second-order inverted pendulum are conducted. Then, an intelligent agent based on the Deep Q-Network(DQN) algorithm is designed for its balance control. To make the control effect more intuitive, a physical simulation model of the pendulum is built using the Simscape Multibody module in Matlab/Simulink. The simulation results show that the agent based on the DQN algorithm can effectively control the balance of the second-order inverted pendulum, proving the effectiveness of the DQN algorithm in the stable control of the pendulum.更多还原