Reinforcement Learning on Quanser Robots

Requirements:

Use command in spec-file.txt to install conda packages.

See Quanser Platform: https://git.ias.informatik.tu-darmstadt.de/quanser/clients

Introduction

Quanser is a simulation platform of physical models. I implement MPC algorithm for tasks on Ball Balance, CartPoleSwing and Qube models.

I apply BeeHive and Random Shooting algorithm for MPC.