様々な椅子での重心・重量による姿勢識別率に関する検討

摘要

ユーザの意図や現在の行動，置かれている状況を取得し，ユーザに適切な支援やサービスを提供しようとする試みとしてコンテキストアウェアネス技術がある．ユーザの状況を考慮できるシステムを構築できれば，状況に応じて有効な情報を選択してユーザに提供できたり，情報提示のタイミングを制御することでユーザの集中を阻害しないようにしたりし，作業効率をより高められると考えられる．%We propose a seated posture recognition technique with no wearable sensor by using seated user's weight and his/her center of balance on the seat to support desk workers attentively. Our previous study indicates this technique can recognize seated postures with fewer non-wearable sensors. However, this recognition technique is only evaluated in unrealistic hardware equipment. In this paper, we evaluate the proposed technique in various kinds of commonly-used chairs which differ from each other in the functions of casters and a rotatable seat. We conduct an experiment to acquire the accuracy of the recognition in these chairs. From the result, we discuss about the useful range of proposed posture recognition technique.