Identifying Multiple Interaction Events from Tactile Data during Robot-Human Object Transfer

TL;DR

This paper demonstrates that tactile sensors can accurately identify multiple interaction events during robot-human object transfer, using machine learning techniques to classify pull, push, bump, or hold actions with high accuracy.

Contribution

The study introduces a method to differentiate multiple interaction events solely through tactile data, employing a Bag of Words feature extraction and SVM classification.

Findings

Over 95% accuracy in classifying four interaction events

Effective feature extraction using Bag of Words approach

Feasibility of tactile-based event detection in robot-human interactions

Abstract

During a robot to human object handover task, several intended or unintended events may occur with the object - it may be pulled, pushed, bumped or simply held - by the human receiver. We show that it is possible to differentiate between these events solely via tactile sensors. Training data from tactile sensors were recorded during interaction of human subjects with the object held by a 3-finger robotic hand. A Bag of Words approach was used to automatically extract effective features from the tactile data. A Support Vector Machine was used to distinguish between the four events with over 95 percent average accuracy.