Applying hybrid reinforcement and unsupervised weightless neural network learning algorithm on autonomous mobile robot navigation

Abstract

An autonomous system constructed using written computer programs based on human expert knowledge only handles anticipated and verified states. On the other hand, a self-learning algorithm allows an autonomous system to instinctively acquire knowledge, learn from experience and be more prepared to expect the unexpected. A novel hybrid self-learning algorithm which combines reinforcement and unsupervised weightless neural network algorithm learning was formulated. The self-learning algorithm was applied to an autonomous mobile robot navigation system in simulation and physical world. The result shows that the simulated and physical robot possesses the ability to self-learn by acquiring knowledge, learn and record experience without having prior information on the environment. The mobile robot was able to distinguish different types of obstacles i.e. corners and walls; and generate complex control sequences of actions in order to avoid these obstacles.