Reinforcement Based Artificial Neural Network

In this paper, we have applied a cognitive based artificial neural network which is used to determine a collision free shortest path of a mobile robot from the initial point to the destination in an unknown environment. In this paper we have created an Artificial Neural Network (ANN) which is used to a path by its nonlinear functional approximation. The training samples of this artificial neural network have been collected by a reinforcement learning method known as Q learning. This path planning algorithm has been devised using five state action mapping relationship. The algorithm was applied on our designed mobile robot manipulator and the results show that the path planning done by this method was better than the path devised out by using ANN and Q learning method separately.