Enhanced Mobile Robot Obstacle Avoidance: A Reinforcement Learning Approach with State-Dependent Action Set

In this paper, we address the challenge of enabling a mobile robot to navigate through environment such as a a hospital corridor while avoiding collisions with obstacles. The main objective of the agent is to navigate through these corridor without colliding with any of the static or dynamic obstacles in its way. The robot or the agent is equipped with three proximity sensors which determines the proximity of the obstacles in it’s respective direction. The learning is achieved through Q-learning algorithm which is a value based reinforcement learning technique. In order to achieve much lesser computational time and cost we implement state dependant actions such that we can decrement the number of iterations for learning. For testing and visualization Python’s Pygame library is involved. The agent undergoes training with a grid-based environment.