Reshaping human intention in Human-Robot Interactions by robot moves A comparative analysis of HMM and OOM methods

Abstract

This paper outlines the methodology and experiments associated with the reshaping of human intentions based on robot movements within Human-Robot Interactions (HRIs). Although studies on estimating human intentions are well studied in the literature, reshaping intentions through robot-initiated interactions is a new significant branching in the field of HRI. In this paper, we analyze how estimated human intentions can intentionally change through cooperation with mobile robots in real Human-Robot environments. This paper proposes an intention-reshaping system that includes either the Observable Operator Models (OOMs) or Hidden Markov Models (HMMs) to estimate human intention and decide which moves a robot should perform to reshape previously estimated human intentions into desired ones. At the low level, the system needs to track the locations of all mobile agents using cameras. We test our system on videos taken in a real HRI environment that has been developed as our experimental setup. The results show that OOMs are faster than HMMs and both models give correct decisions for testing sequences.