An autonomous robot: Using ANN to navigate in a static path

Abstract

A robot with autonomous navigation is not only able to find and follow its exact path but also able to avoid an obstacle which comes in its way without human assistance. This paper deals with an intelligent control of an autonomous robot which is trained with Artificial Neural Network to navigate in a partially structured environment which is full of static obstacles. The training capability according to the sensory input and its response to the obstacles are focused as two main challenges. In this work a Neural Network model is developed. This is done by algorithm of Artificial Intelligence. Then it is trained in Arduino platform for a navigation system designed for autonomous robot. To train the robot, a number of training samples are introduced. Ultrasonic sonar sensors are used with the Neural Network in order to find its route without colliding with any obstacle after the training. This novel approach in robot navigation is expected to open new doors of artificial intelligence in future.