부산대학교 도서관

This paper presents a novel approach for personal robots to identify and track the walking owner, by using an omnidirectional camera on the robot and IR LED tags emitting particular lighting patterns. For the ageing society, personal robots which bring baggage and follow the owner, can improve the quality of life of elderly people. Robust identification of the owner is one of the important issues for this purpose. In this paper, a new pattern classification method based on the particle filter is proposed to identify the light emitting tags on the owner's feet. Proposed method estimates the cycle and phase of the lighting patterns as to track the tags robustly even in outdoor environment. The effectiveness of the proposed approach is verified by some experiments using our robot and prototype of the tag.