Perceptions or Actions? Grounding How Agents Interact Within a Software Architecture for Cognitive Robotics

Rebeca Marfil, Adrian Romero-Garces, Juan P. Bandera, Luis J. Manso, Luis V. Calderita, Pablo Bustos, Antonio J. Bandera, Javier Garcia-Polo, Fernando Fernandez, Dimitri Voilmy

Research output: Contribution to journalArticlepeer-review

Abstract

One of the aims of cognitive robotics is to endow robots with the ability to plan solutions for complex goals and then to enact those plans. Additionally, robots should react properly upon encountering unexpected changes in their environment that are not part of their planned course of actions. This requires a close coupling between deliberative and reactive control flows. From the perspective of robotics, this coupling generally entails a tightly integrated perceptuomotor system, which is then loosely connected to some specific form of deliberative system such as a planner. From the high-level perspective of automated planning, the emphasis is on a highly functional system that, taken to its extreme, calls perceptual and motor modules as services when required. This paper proposes to join the perceptual and acting perspectives via a unique representation where the responses of all software modules in the architecture are generalized using the same set of tokens. The proposed representation integrates symbolic and metric information. The proposed approach has been successfully tested in CLARC, a robot that performs Comprehensive Geriatric Assessments of elderly patients. The robot was favourably appraised in a survey conducted to assess its behaviour. For instance, using a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree), patients reported an average of 4.86 when asked if they felt confident during the interaction with the robot. This paper proposes a mechanism for bringing the perceptual and acting perspectives closer within a distributed robotics architecture. The idea is built on top of the blackboard model and scene graphs. The modules in our proposal communicate using a short-term memory, writing the perceptual information they need to share with other agents and accessing the information they need for determining the next goals to address.
Original languageEnglish
Pages (from-to)479–497
Number of pages27
JournalCognitive Computation
Volume12
Early online date9 Nov 2019
DOIs
Publication statusPublished - Mar 2020

Bibliographical note

© Springer Nature B.V. 2019. The final publication is available at Springer via http://dx.doi.org/10.1007/s12559-019-09685-5

Keywords

  • Automatic planning
  • Cognitive robotics
  • Software architectures

Fingerprint

Dive into the research topics of 'Perceptions or Actions? Grounding How Agents Interact Within a Software Architecture for Cognitive Robotics'. Together they form a unique fingerprint.

Cite this