Signal Communication for Collaborative Robot Behavior Control

Abstract

Robots are increasingly required to work in close cooperation with humans and other robots, performing common tasks jointly, in collaboration. High-level decision making by an autonomous agent with such interactions require specific means of responding to the situation. This paper describes the implementation of the control system for a collaborative robot, based on the emotion-need architecture that provides reliable and interpretable interaction means of the robot with other agents. The robot is an autonomous device whose behavior is aimed at satisfying its current needs. The communication between the operator and the robot is based on signal communication: the operator’s signal ("command") activates an element of the emotion-need scheme — a gateway that implements the emotional feedback. Since the gateway connects the sensors and behavioral procedures, its activation starts the corresponding procedure, as if the sensory system had registered a real stimulus. The external signal of such indirect control can be represented in acoustic form or as the operator’s posture. The needs of the robot are aimed at both its physical survival (danger avoidance) and maintenance of the social connections (accompanying friendly agents). The robot recognizes external commands and activates the gateway associated with the highest priority action. The activity also gradually fades, which allows actions to be performed for some time even in the absence of an appropriate internal stimulus. The resulting robot is able to perform the simplest patrol tasks together with a human, acting in a predictable and understandable (interpretable) way for the human.