Display options
Cite
Jauffret A, Cuperlier N, Tarroux P, et al. From self-assessment to frustration, a small step toward autonomy in robotic navigation. Front Neurorobot. 2013;7:16doi: 10.3389/fnbot.2013.00016.
Jauffret, A., Cuperlier, N., Tarroux, P., & Gaussier, P. (2013). From self-assessment to frustration, a small step toward autonomy in robotic navigation. Frontiers in neurorobotics, 716. https://doi.org/10.3389/fnbot.2013.00016
Jauffret, Adrien, et al. "From self-assessment to frustration, a small step toward autonomy in robotic navigation." Frontiers in neurorobotics vol. 7 (2013): 16. doi: https://doi.org/10.3389/fnbot.2013.00016
Jauffret A, Cuperlier N, Tarroux P, Gaussier P. From self-assessment to frustration, a small step toward autonomy in robotic navigation. Front Neurorobot. 2013 Oct 08;7:16. doi: 10.3389/fnbot.2013.00016. eCollection 2013. PMID: 24115931; PMCID: PMC3792359.
Copy Download .nbib Format: NLM
Share it on

Front Neurorobot. 2013 Oct 08;7:16. doi: 10.3389/fnbot.2013.00016. eCollection 2013.

From self-assessment to frustration, a small step toward autonomy in robotic navigation.

Frontiers in neurorobotics

Adrien Jauffret, Nicolas Cuperlier, Philippe Tarroux, Philippe Gaussier

Affiliations

  1. Neurocybertic Team, Equipes Traitement de l'Information et Systèmes Laboratory, UMR 8051 Cergy, France.

PMID: 24115931 PMCID: PMC3792359 DOI: 10.3389/fnbot.2013.00016

Abstract

Autonomy and self-improvement capabilities are still challenging in the fields of robotics and machine learning. Allowing a robot to autonomously navigate in wide and unknown environments not only requires a repertoire of robust strategies to cope with miscellaneous situations, but also needs mechanisms of self-assessment for guiding learning and for monitoring strategies. Monitoring strategies requires feedbacks on the behavior's quality, from a given fitness system in order to take correct decisions. In this work, we focus on how a second-order controller can be used to (1) manage behaviors according to the situation and (2) seek for human interactions to improve skills. Following an incremental and constructivist approach, we present a generic neural architecture, based on an on-line novelty detection algorithm that may be able to self-evaluate any sensory-motor strategies. This architecture learns contingencies between sensations and actions, giving the expected sensation from the previous perception. Prediction error, coming from surprising events, provides a measure of the quality of the underlying sensory-motor contingencies. We show how a simple second-order controller (emotional system) based on the prediction progress allows the system to regulate its behavior to solve complex navigation tasks and also succeeds in asking for help if it detects dead-lock situations. We propose that this model could be a key structure toward self-assessment and autonomy. We made several experiments that can account for such properties for two different strategies (road following and place cells based navigation) in different situations.

Keywords: action selection; bio-inspired robotics; metalearning; neural-networks; self-assessment; sensory-motor system

References

  1. Am Psychol. 1989 Feb;44(2):112-9 - PubMed
  2. Biol Cybern. 2001 Jun;84(6):401-10 - PubMed
  3. Child Dev. 2000 Jul-Aug;71(4):1061-71 - PubMed
  4. Nature. 1996 Sep 19;383(6597):256-9 - PubMed
  5. Biol Cybern. 2002 Jan;86(1):15-28 - PubMed
  6. Neural Comput. 2005 Jun;17(6):1339-84 - PubMed
  7. Monogr Soc Res Child Dev. 1992;57(1):1-98 - PubMed
  8. Bioinspir Biomim. 2012 Jun;7(2):025009 - PubMed
  9. Neural Netw. 2013 Jul;43:8-21 - PubMed
  10. Biol Cybern. 1977 Aug 3;27(2):77-87 - PubMed
  11. Behav Brain Sci. 2005 Apr;28(2):169-94; discussion 194-245 - PubMed
  12. Front Neurorobot. 2007 Nov 02;1:3 - PubMed
  13. Hippocampus. 2001;11(5):551-68 - PubMed
  14. Biol Cybern. 1976 Jan 8;21(2):85-95 - PubMed

Publication Types