Autonomous robot behaviours for co-operative agents using fuzzy logic and subtractive clustering

Intelligent autonomous robots and multiagent systems, having different skills and capabilities for specific subtasks, have the potential to solve problems more efficiently and effectively. In this paper both f i m y logic (FL) and subtractive clustering (SC) are used for the design of autonomous robot behaviours. The design procedure is conducted in two stages: first subtractive clustering is applied to extract fuzzy model from experimental data; then adaptive neuro-fuzzy inference system (ANFIS) is applied to improve the fuzzy model performance. This technique produces good result (0.01% root mean square error) and has the advantage of being closer to natural human language, by describing the robot behaviours using a set of linguistic rules.