Feed-forward Neural Network approach in Navigation control and P-I speed control of Mobile Robot

Abstract: In this article, the implementation of Artificial Neural Network (ANN) approach for obstacle avoidance in Mobile Robots is presented. A basic method to interpret the data from the proximity sensors and a single layer feed forward Neural Network algorithm for the navigation control is proposed. The proximity sensor array outputs are arranged in combinations of ‘0’ and ‘1’. This strategy allows training the Neural Network using only one type of sensor to detect the presence of obstacle. A suitable algorithm is applied for adjusting the network weights. User directed the robot through a GUI. In addition to this, the speed control of Mobile robot is achieved by a Proportional- Integral controller. The trained neural network has been tested with the real sensor data. The robot has been wirelessly connected with the Neural Network controller. Experimental results validate the effectiveness of the proposed approach in mobile robot navigation. The training of Neural Network is carried out by using MATLAB platform. Keywords: Artificial Neural Network; Navigation Control; PI; Mobile Robot; MATLAB. Title: Feed-forward Neural Network approach in Navigation control and P-I speed control of Mobile Robot Author: Jeffina Stephens S, Maheshan C M, Prasanna Kumar H International Journal of Electrical and Electronics Research ISSN 2348-6988 (online) Research Publish Journals

Citation