Robotic Arm

About

Time: Jan 2022~ May 2022

Skills Gained: Reinforcement Learning, Model predictive control, Pytorch

Target

Allowed robotic arm's end effector to stop by the given point within 2.5 seconds with speed less than 0.05m/s and distance less than 0.01m. We will used three method to perform this tasks.

I have tried out three following methods:

Method 1: Model Predictive Control (MPC)

Method 2:Deep Reinforcement Learning (DQN)

Method 3: Proximal Policy Optimization (PPO)