A low cost microcontroller implementation of neural network based hurdle avoidance controller for a car-like robot

Abstract

This paper describes the implementation of a neural network based hurdle avoidance controller for a car like robot using a low cost single chip 89C52 microcontroller. The neural network is the multilayer feed-forward network with back propagation training algorithm. The network is trained offline with tangent-sigmoid as activation function for neurons and is implemented in real time with piecewise linear approximation of tangent-sigmoid function. Results have shown that up-to twenty neurons in hidden layer can be deployed with the proposed technique using a single 89C52 microcontroller. The vehicle is tested in various environments containing obstacles and is found to avoid obstacles in its path successfully.