Optimal robot path planning system by using a neural network-based approach

This paper proposes an optimal robot path planning system that can build map, plan optimal paths, and maneuver mobile robots. The system constructs a grid-based map by using information on the locations of the origin and static obstacles. The system calculates the optimal trajectory by using a simplified neural network model and accordingly maneuvers a mobile robot. For dynamic obstacles, the mobile robot can sense the ambient environment and avoid possible collisions. A practical experiment using an Arduino-based platform was conducted to illustrate the effectiveness of the proposed methodology.