A multi-expert model for dialogue and behavior control of conversational robots and agents

Abstract

This paper presents an intelligence model for conversational service robots. It employs modules called experts, each of which is specialized to execute certain kinds of tasks such as performing physical behaviors and engaging in dialogues. Some of the experts take charge in understanding human utterances and deciding robot utterances or actions. The model enables switching and canceling tasks based on recognized human intentions, as well as parallel execution of several tasks. This model specifies the interface that an expert must have, and any kind of expert can be employed if it conforms to the interface. This feature makes the model extensible.Key words: conversational robot, conversational agent, robot intelligence, behavior and dialogue control, multi-expert model ⋆ This paper is a considerably extended version of [1]. We thank the Association for Computational Linguistics for permission to reuse the material.

A Stimulus-Response Framework for Robot Control

“…Only recently has the need to model the stimulus-response pattern leading to task switching and the cognitive functions of shifting and inhibition become a hot topic in cognitive robotics [Medina Ayala et al 2012;Nakano et al 2011;Ganesh and Burdet 2013;Ajoudani et al 2013;Ding and Fang 2013;Tao et al 2012;Li et al 2013;Sung et al 2013;Jamone et al 2013;Saglam and Byl 2013;and Li and Cheah 2012]. The earliest studies were carried out within brain-actuated interaction [Milln et al 2004], mechatronics [Capi 2007], behavior learning [Ito et al 2006], navigation [Althaus and Christensen 2003], and planning [Finzi and Pirri 2005].…”

“…The need to model adaptation, to increase the flexibility in handling complex tasks, has led to modeling task switching for robot navigation, motion planning, and control in large-scale environments [Althaus and Christensen 2003], for path planning [Medina Ayala et al 2012], for recognizing human intentions in conversation [Nakano et al 2011], and for motor planning to improve tasks execution [Ganesh and Burdet 2013]. Task switching is also addressed in manipulation [Ajoudani et al 2013;Ding and Fang 2013] and soft robot control [Tao et al 2012;Li et al 2013], for modeling multirobot systems [Sung et al 2013], for learning task space control through the use of different tools [Jamone et al 2013], for biped walking in rough terrain [Saglam and Byl 2013], and for multiple-space control [Li and Cheah 2012].…”

Communicative commitments: Model checking and complexity analysis

“…In [36], the authors present a general model for controlling the behavior of autonomous robots/agents that can engage in a conversational dialog with humans. The core components of this model are called ''experts'', which are responsible for understanding humans' requests (in general, humans' utterances), providing useful information, and deciding on robots actions.…”

An active audition framework for auditory-driven HRI: Application to interactive robot dancing

“…Auditory-driven interactive robotic systems range from musical [3] and dancing robots [4] to conversational agents [5]. These are applied in different social contexts such as entertainment, pedagogical, and therapeutic scenarios.…”