Experimental Study on Reinforcement Learning-based Control of an Acrobot
This work provides insights into applying RL for controlling an Acrobot, which is relevant for robotics and energy harvesting applications, representing an incremental step in control systems.
This paper explores the use of reinforcement learning (RL) to control an Acrobot, focusing on regulating its angular velocity and total energy. The RL algorithm successfully drives the angular velocity or energy of the first pendulum towards desired values, achieving libration or full rotation of the unactuated pendulum.
We present computational and experimental results on how artificial intelligence (AI) learns to control an Acrobot using reinforcement learning (RL). Thereby the experimental setup is designed as an embedded system, which is of interest for robotics and energy harvesting applications. Specifically, we study the control of angular velocity of the Acrobot, as well as control of its total energy, which is the sum of the kinetic and the potential energy. By this means the RL algorithm is designed to drive the angular velocity or the energy of the first pendulum of the Acrobot towards a desired value. With this, libration or full rotation of the unactuated pendulum of the Acrobot is achieved. Moreover, investigations of the Acrobot control are carried out, which lead to insights about the influence of the state space discretization, the episode length, the action space or the mass of the driven pendulum on the RL control. By further numerous simulations and experiments the effects of parameter variations are evaluated.