PDRplus: Human Behaviour Sensing Method for Service Field Analysis
This chapter presents a novel method of estimating position, orientation, and multiple actions of a worker in a service field. In general, pedestrian dead reckoning (PDR) is appropriate for effectively estimating the position and orientation of a pedestrian in an indoor environment. However, in actual service fields, PDR is not as accurate for workers’ behaviour sensing when a number of actions for their work other than walking are taking place. Moreover, common sensors for PDR have less information for multiple action recognition other than walking. For realizing human behaviour sensing for service process analysis, we propose a method which integrates human localization and action recognition with the complementary use, named “PDRplus”. In service fields, since position, orientation, and action of a human usually show strong correlation with her or his situation, both the PDR and action recognition can be improved with complementary use of the PDR and action recognition. In this chapter, in order to ensure the effect of the complementary use of the PDR and action recognition, we conducted two types of experiments in real service industry fields. Firstly, we compared accuracies of the action recognition both with and without using the PDR in the restaurant kitchen, and average recognition rate of five types of actions was improved about 19 % points. Secondly, we compared accuracies of the PDR both with and without using the action recognition in house-assembly plants, and average position error was reduced by 19.5 %.
KeywordsAction recognition Boosting Pedestrian dead reckoning Service engineering
This work was supported by the Ministry of Economy, Trade and Industry (METI) of Japan.
- 1.Imai M (1986) Kaize: the key to Japan’s competitive success. McGraw-Hill, New YorkGoogle Scholar
- 3.Woodman O, Harle R (2008) Pedestrian localisation for indoor environments. In: Proceedings of the 10th international conference on ubiquitous computing (UbiComp), ACM Press, 344:114–123Google Scholar
- 4.Judd T, Vu T (2008) Use of a new pedometric dead reckoning module in GPS denied environments. In: Proceedings of the IEEE/ION position, location and navigation (PLANS2008), pp 120–128Google Scholar
- 5.Kourogi M, Kurata T (2003) Personal positioning based on walking locomotion analysis with self-contained sensors and a wearable camera. In: Proceedings of the ISMAR2003, pp 103–112Google Scholar
- 6.Kourogi M, Ishikawa T, Kurata T (2010) A method of Pedestrian dead reckoning using action recognition. In: Proceedings of the IEEE/ION position, location and navigation symposium (PLANS), pp 85–89Google Scholar
- 7.Lester J, Choudhury T, Borriello G (2006) A practical approach to recognizing physical activities. In: Proceedings of the pervasive’06, pp 1–16Google Scholar
- 8.Bao L, Intille SS (2004) Activity recognition from user-annotated acceleration data. In: Proceedings of the 2nd international conference on pervasive, computing, pp 1–17Google Scholar
- 9.Jin Y, Toh HS, Soh WS, Wong WC (2011) A robust dead-reckoning Pedestrian tracking system with low cost sensors. In: Proceedings of the IEEE international conference on pervasive computing and communications (Percom), pp 222–230Google Scholar
- 10.Freund Y, Schapire RE (1995) A decision-theoretic generalization of on-line learning and application to boosting. Computational learning theory: eurocolt, pp 23–37Google Scholar