Designing an artificial attention system for social robots

Abstract

In this paper, we introduce the main components comprising the action-perception loop of an overarching framework implementing artificial attention, designed to fulfil the requirements of social interaction (i.e., reciprocity, and awareness), with strong inspiration on current theories in functional neuroscience. We demonstrate the potential of our framework, by showing how it exhibits coherent behaviour without any inbuilt prior expectations regarding the experimental scenario. Current research in cognitive systems for social robots has suggested that automatic attention mechanisms are essential to social interaction. In fact, we hypothesise that enabling artificial cognitive systems with middleware implementing these mechanisms will empower robots to perform adaptively and with a higher degree of autonomy in complex and social environments. However, this type of assumption is yet to be convincingly and systematically put to the test. The ultimate goal will be to test our working hypothesis and the role of attention in adaptive, social robotics.

How to Build a Supervised Autonomous System for Robot-Enhanced Therapy for Children with Autism Spectrum Disorder

“…Some simplifications of the model were applied to adapt it for our context. Other approaches like [78] were taken into account. This approach merges top-down and perceptual attention in an efficient manner.…”

Object Detection and Recognition for Assistive Robots: Experimentation and Implementation

“…saliency) attentional capture by targets versus distractors. In this regard, a four-module attentional architecture has been defined by Lanillos et al [45] in which the first module corresponds to the perception sense by building an egocentric map according to relevance encoded as saliency. This information is fed to the top-down controller which ensures that the selection of the new focus of attention will take into account the current system goals and context.…”

Gaze Tracing in a Bounded Log-Spherical Space for Artificial Attention Systems

“…In this sense, the direction of other's attention is crucial for mastering social interaction [16]. In fact, in the case of joint attention [3,7], where two individuals perform a triadic relationship between them and an object, gaze direction provides useful information to know if the human is looking at the robot as well as about the direction of the shared object [2]. This location is potentially important for the correct social interaction and can also play a significant role in the task being effectuated.…”

“…In the study presented in [7], where an artificial attention system is tested without integrating the gaze interpretation ability, it is clear that the robot cannot infer the complete non-verbal message intended by the interlocutor. In that experiment, the interlocutor shows an object to the robot by deictic fixation (e.g.…”

“…In this paper we present a discrete log-spherical gaze tracing algorithm to be integrated in a full-fledged attention system [7] that establish a correspondence between a set of discretised cells with potential objects being fixated by the observer. We integrate the egocentric representation of the environment, Bayesian Volumetric Map (BVM) [4], with the gaze cue, improving the capacity of the machine to infer important regions even if they are outside the field of view and providing the robot with the representational stage, thus laying the foundation for non-verbal social HRI.…”