Supervised Learning-based Multi-Modal Perception for Robot Partners using Smart Phones

Abstract

This paper proposes a method for multi-modal perception of human-friendly robot partners based on various types of sensors built in a smart phone. The proposed method can estimate human interaction modes by fuzzy spiking neural network. The learning method of the spiking neural network based on the time series of the measured data is explained as well. Evolution strategy is used for optimizing the parameters of the fuzzy spiking neural network. Several experimental results are presented for confirming the effectiveness of the proposed technique. Finally, the future direction on this research is discussed.