Intelligent Path Planning for Robots and Practical Implementation of Programmable Headlights for Autonomous Vehicles

The ability to move is essential for the development of intelligent robots for autonomous navigation.Neural networks outperform traditional methods in modeling complex relationships and identifying patterns, but current systems are limited to specific robots and sensors.This paper presents a universal method for interpreting data from different 2D sensors, predicting distances between robots and walls, and using neural networks for navigation.The goal is to create a versatile algorithm that can be applied to different robots and programmable lamps, reducing accidents.The thesis also explores programmable lamps that block light from reaching the eyes of passengers, using one network to determine free space using odometry data and another to find safe paths while avoiding obstacles.Simulated path examples will be presented.