I am trying to teach a double pendulum fully actuated to perform a swing-up maneuver and keep the position if it reaches it. I previously trained a single pendulum with DQN and it learned the policy well in just a couple of hours of training. Now the problem is that I have a robot with two joints, so two actions must be chosen at the same time: the torque for the first one and the torque for the second one. The only method that comes to me is to generate the Q value for each possible action pair. The problem with this is that the action space grows exponentially with the joint space. Is there any other way to solve the problem?
Asked
Active
Viewed 158 times
