Robots and controls systems are increasingly adopting Intelligent non-linear-based control scheme approaches. Such non-linear systems become a complex computational problem for linear control schemes due to the rigorous calculation. Since most systems are inherently non-linear, following through with linear control schemes becomes an exhaustive process. In this research work, we wish to present reinforcement learning-based control with PPO (Proximal Policy Optimization) and MPC (Model Predictive Control) control schemes on our chosen assembly, i.e., the wheeled robot. We will be comparing these control schemes against each other based on the efficiency and performance of these systems for a specified task, including point-to-point locomotion in 3D space.

No abstract

No abstract