header

Information and Sciences

Classification of Age Groups Using Walking Pattern Obtained from a Laser Range Scanner

Professor

Yuka KATO, Ph.D.

Department of Information and Sciences


Why

Due to the spread of communication robots, much attention has been paid to researches on human-robot interaction. To realize a society of human-robot coexistence, it is important for robots to recognize their working environment including human properly and to provide their services according to the situation.

How

We proposed a classification method of age groups (elderly and young people) based on the Support Vector Machine (SVM), which is a machine learning model, by using the movement data of people's legs obtained from a Laser Range Scanner (LRS) We constructed a predictor for the classification by using 25 samples.



Findings

We classified the elderly from the young with the accuracy of 80-90%. We also showed that the classification accuracy improved particularly by using acceleration and standard deviation as the feature values.

Papers

ーS. Sakai, S. Kimura, D. Nomiyama, T. Ikeda, N. Matsuhira and Y. Kato, "Classification of Age Groups Using Walking Data Obtained from a Laser Range Scanner", Proc. on IECON 2016, pp.5862-5867, 2016.

ーS. Sakai, S. Kimura, D. Nomiyama, T. Ikeda, N. Matsuhira and Y. Kato, "A Classification Method of Elderly and Young People Using Walking Pattern Obtained from a Laser Range Scanner", IPSJ Journal, Vol.58, No.2, pp.375-383, 2017. (Specially Selected Paper)



Professor

Yuka KATO, Ph.D.

Department of Information and Sciences

My research fields are information networks, ubiquitous computing and robot services using computer networks. In our laboratory, for example, we are working on the following research topics:
・Remote navigation of mobile robots using cloud computing environments
・A pedestrian model in human-robot coexisting environments for mobile robot navigation
・Property recognition of a person for human-robot interaction