Multi-Robot Behavior Adaptation to Humans' Intention in Human-Robot Interaction Using Information-Driven Fuzzy Friend-Q Learning.

Lue-Feng Chen,Zhen-Tao Liu,Min Wu,Fangyan Dong,Kaoru Hirota

DOI: https://doi.org/10.20965/jaciii.2015.p0173

2015-01-01

Journal of Advanced Computational Intelligence and Intelligent Informatics

Abstract:A multi-robot behavior adaptation mechanism that adapts to human intention is proposed for human-robot interaction (HRI), where information-driven fuzzy friend-Q learning (IDFFQ) is used to generate an optimal behavior-selection policy, and intention is understood mainly based on human emotions. This mechanism aims to endow robots with human-oriented interaction capabilities to understand and adapt their behaviors to human intentions. It also decreases the response time (RT) of robots by embedding the human identification information such as religion for behavior selection, and increases the satisfaction of humans by considering their deep-level information, including intention and emotion, so as to make interactions run smoothly. Experiments is performed in a scenario of drinking at a bar. Results show that the learning steps of the proposal is 51 steps less than that of the fuzzy production rule based friend-Q learning (FPRFQ), and the robots' RT is about 25% of the time consumed by FPRFQ. Additionally, emotion recognition and intention understanding achieved an accuracy of 80.36% and 85.71%, respectively. Moreover, a subjective evaluation of customers through a questionnaire obtains a reaction of "satisfied." Based on these preliminary experiments, the proposal is being extended to service robots for behavior adaptation to customers' intention to drink at a bar.