Development of Machine Learning Algorithm for Ground Vehicle Control

Motivation: In this research, we develop an algorithm based on machine learning for self-driving ground vehicle robots. Visual feedback based on Machine learning is employed as the main algorithm for self-driving robot

Open topics: Two-wheeled mobile robot using LQR, using MPC, and Reinforcement Learning

H-Infinity Controller Design of Two-Wheeled Mobile Robot Under Disturbance

This study investigates a controller design on a two-wheeled robot model under disturbance. The concept of a two-wheeled robot is based on the principle of an inverted pendulum model. The control is investigated on the state-space-based controller design. The controller is designed to satisfy a two-wheeled self-balancing robot's stability, capable of maintaining the robot in a steady vertical position and robust to disturbance. The main focus of this study is developing the H-infinity controllers for the two-wheeled robot model. The controller is simulated through trial and error and compared to distinguish the best performing controller from other Linear Quadratics Controller (LQR) methods to achieve a solid validation. This investigation shows that the H-infinity controller performs a better performance in the two-wheeled robot model under disturbance