Rapid learning of sensor-based behaviours of mobile robots based on B-spline fuzzy controllers

Abstract

We present a robust and safe approach for learning of sensor-based controllers and show how it can make the programming of control tasks easier It is first discussed how to construct a fuzzy controller with B-splines and why rapid convergence of its learning can be achieved. To apply the concept in learning of primitive behaviours of mobile robots like collision avoidance with different environments and tracking object contours, we propose several gradual steps of active online learning. We point out that rapid convergence of this learning procedure can be guaranteed, which is confirmed by diverse real experiments with a mobile robot which shows these behaviours in an arbitrary environment after only a few learning steps.