Adaptive Behavior Control for Robot Soccer Navigation Using Fuzzy-based Social Force Model

Abstract

Navigation is one of the capabilities that any mobile robot must-have when moving from one position to another. How to move effectively becomes crucial when a navigation skill must be practiced by a mobile robot soccer in a dynamic environment with high speed. This paper proposes the use of the Fuzzy-based Social Force Model (F-SFM) to control the navigation of a soccer robot. In the framework of the Social Force Model (SFM), the speed and direction of navigation are determined by the resultant force calculating from the attractive force by the goal and the repulsive force by obstacles. The amount of repulsive force generated by each obstacle is largely determined by the stimulus received by the robot and the pre-defined SFM parameters setting. Here, the Fuzzy Inference System (FIS) is used to adapt to one SFM parameter, i.e., gain factor, k, based on the stimulus received, namely the relative distance of the obstacle, d, and its direction, γ. With the ability of parameter to change adaptively, the reactivity and responsiveness of the robot can be controlled. Based on the experimental results using a realistic 3D simulator V-Rep, the implementation of adaptive SFM parameter values outperformed the constant one.