See what i mean-Probabilistic optimization of robot pointing gestures

Abstract

Humans use gestures such as pointing extensively in order to anchor linguistic expressions to objects in the physical world. Similarly gestures can be valuable in decentralized robotic systems, allowing communication between agents and transfer of symbolic meanings. Pointing gestures are especially valuable in crowded scenes where multiple possible matches are present. However, pointing in crowded scenes can itself remain ambiguous if the pointing direction is not carefully chosen. This paper proposes a probabilistic model for pointing and gesture detection accuracy. The model allows planning optimal pointing actions by minimizing the probability of pointing errors due to ambiguities and limited accuracy. We also describe how to measure the accuracy of an agent's pointing gesture and to calibrate the model for that agent. Experimental results suggest that the proposed model captures the qualitative behavior of pointing success well.