A Learning Algorithm to Select Consistent Reactions to Human Movements



A balance between adaptiveness and consistency is desired for a robot to select control laws to generate reactions to human movements. Learning algorithms are usually employed for the robot to predict the human actions, and then select appropriate reactions accordingly. Two popular classes of learning algorithms, the weighted majority algorithms and the online Winnow algorithms, are biased for either strong adaptiveness or strong consistency. The dual expert algorithm (DEA), proposed in this chapter, is able to achieve a tradeoff between adaptiveness and consistency. We give theoretical analysis to rigorously characterize the performance of the DEA. Both simulation results and experimental data are demonstrated to confirm that DEA enables a robot to learn the preferred reaction to pass a human in a hallway setting. The results may be generalized to other types of human–robot collaboration tasks.


Learning Reaction to human movements Adaptiveness Consistency 


  1. 1.
    Huttenrauch H, Green A, Norman M, Oestreicher L, Eklundh KS (2004) Involving users in the design of a mobile office robot. IEEE Trans Syst Man Cybern Part C: Appl Rev 34(2):113–124Google Scholar
  2. 2.
    Mutlu B, Forlizzi J (2008) Robots in organizations: the role of workflow, social, and environmental factors in human-robot interaction. In: 3rd ACM/IEEE international conference on human-robot interaction (HRI). IEEE, pp 287–294Google Scholar
  3. 3.
    Breazeal C (2004) Social interactions in hri: the robot view. IEEE Trans Syst Man Cybern Part C: Appl Rev 34(2):181–186Google Scholar
  4. 4.
    Dautenhahn K, Woods S, Kaouri C, Walters ML, Koay KL, Werry I (2005) What is a robot companion-friend, assistant or butler? In: IEEE/RSJ international conference on intelligent robots and systems. IEEE, pp 1192–1197Google Scholar
  5. 5.
    Kuncheva L (2004) Classifier ensembles for changing environments. In: Roli F, Kittler J, Windeatt T (eds) Multiple classifier systems, ser. Lecture notes in computer science, vol 3077. Springer Berlin, Heidelberg, pp 1–15Google Scholar
  6. 6.
    Polikar R (2012) Ensemble learning. In: Zhang C, Ma Y (eds) Ensemble machine learning. Springer, US, pp 1–34Google Scholar
  7. 7.
    Littlestone N, Warmuth MK (1989) The weighted majority algorithm. In: (1989) 30th annual symposium on foundations of computer science, vol 1989. IEEE, pp 256–261Google Scholar
  8. 8.
    Littlestone N (1988) Learning quickly when irrelevant attributes abound: a new linear-threshold algorithm. Mach Learn 2(4):285–318Google Scholar
  9. 9.
    Blum A (1997) Empirical support for winnow and weighted-majority algorithms: results on a calendar scheduling domain. Mach Learn 26(1):5–23Google Scholar
  10. 10.
    Freund Y, Schapire RE (1997) A decision-theoretic generalization of on-line learning and an application to boosting. J Comput Syst Sci 55(1):119–139Google Scholar
  11. 11.
    Freund Y (2001) An adaptive version of the boost by majority algorithm. Mach Learn 43(3):293–318Google Scholar
  12. 12.
    Tsymbal A (2004) The problem of concept drift: definitions and related work. Computer Science Department, Trinity College Dublin 106:Google Scholar
  13. 13.
    Kolter JZ, Maloof MA (2007) Dynamic weighted majority: an ensemble method for drifting concepts. J Mach Learn Res 8:2755–2790Google Scholar
  14. 14.
    Crammer K, Mansour Y, Even-Dar E, Vaughan JW (2010) Regret minimization with concept drift. In: COLT, pp 168–180Google Scholar
  15. 15.
    Ross GJ, Adams NM, Tasoulis DK, Hand DJ (2012) Exponentially weighted moving average charts for detecting concept drift. Pattern Recognit Lett 33(2):191–198Google Scholar
  16. 16.
    Pacchierotti E, Christensen HI, Jensfelt P (2006) Design of an office-guide robot for social interaction studies. In: IEEE/RSJ international conference on intelligent robots and systems. IEEE, pp 4965–4970Google Scholar
  17. 17.
    Pacchierotti E, Christensen HI, Jensfelt P (2006) Evaluation of passing distance for social robots. In: The 15th IEEE international symposium on robot and human interactive communication. IEEE, pp 315–320Google Scholar
  18. 18.
    Lauckner M, Kobiela F, Manzey D (2014) Hey robot, please step back!-exploration of a spatial threshold of comfort for human-mechanoid spatial interaction in a hallway scenario. In: The 23rd IEEE international symposium on robot and human interactive communication. IEEE, pp 780–787Google Scholar
  19. 19.
    Severinson-Eklundh K, Green A, Hüttenrauch H (2003) Social and collaborative aspects of interaction with a service robot. Robot Auton Syst 42(3):223–234Google Scholar
  20. 20.
    Chung S-Y, Huang H-P (2010) A mobile robot that understands pedestrian spatial behaviors. In: IEEE/RSJ international conference on intelligent robots and systems (IROS), pp 5861–5866Google Scholar
  21. 21.
    Chung S-Y, Huang H-P (2010) Predictive navigation by understanding human motion patterns. Int J Adv Robot Syst 8(1), 52–64Google Scholar
  22. 22.
    Sisbot EA, Marin-Urias LF, Alami R, Simeon T (2007) A human aware mobile robot motion planner. IEEE Trans Robot 23(5):874–883Google Scholar
  23. 23.
    Young C, Zhang F (2016) A learning algorithm to select consistent reactions to human movements. In: Proceedings of 2016 American control conferences, pp 6152–6157Google Scholar

Copyright information

© Springer International Publishing Switzerland 2017

Authors and Affiliations

  1. 1.School of Electrical and Computer EngineeringGeorgia Institute of TechnologyAtlantaUSA

Personalised recommendations