Mobile robot path-learning to separate goals on an unknown world

摘要

In this article we face the problem of navigating a mobile robot on an unknown indoor environment. The parti-game approach is used for simultaneous learning of a world model, and learning a path from an initial location to a specified goal region. These two learning abilities may be seen as cooperating and enhancing each other in order to improve the overall system performance. It is shown that the constructed world model is general-purpose, in the sense that its usefulness is not restricted to be used on self-learning a particular path, but may be valuable for learning paths with different (start, goal) pairs. The robot uses its own infrared distance-sensors to perform obstacle detection while moving. Is also has the predefined ability of performing straight-line motions. Simulation results are presented that validate the effectiveness of the approach.