Efficacy of gesture for communication among humanoid robots by fuzzy inference method

Abstract

Social robots can make use of communication in various environments such as space, nuclear plants, and hospitals, etc., where they are expected to communicate with other robots or with human beings. Imitation learning for behaviour-based programming is exploited and used for robot teaching. In this paper, we achieve gesture-based communication with fuzzy controller between two HOAP-2 robots. The approach uses the power of rule-based system and inference engine to recognise and generate the interactive gestures. It filters the noise in recognition process as well as it also avoids the ambiguity of multiple solutions during gesture generation phase. We demonstrate the validity of this approach through kickboxing gestures of HOAP-2 robots, where they can respond to each other's actions.