| 菅田 唯仁 | M, 2回目発表 | ユビキタスコンピューティングシステム | 安本 慶一, 岡田 実, 荒川 豊, 藤本 まなと |
|
title: Battery-less place recognition system using multiple type solar cells
abstract: We propose a room-level place recognition system which uses energy harvesting elements as both sensor and power source. We succeeded to distinguish ten places thanks to different characteristics of solar cells according to their materials. However, since the previous research focused on investigating the feasibility of our fundamental idea, it used a microcomputer board that requires an additional battery to measure the amount of generated power. Hence, it did not reach the level to utilize the energy harvesting elements as a power supply required for running the whole system. In this study, we re-design our previously proposed mechanism to measure the amount of generated power of the energy harvesting elements and propose a novel mechanism to measure the amount of generated power by using the generated power without any additional battery. We designed and implemented the new circuit on PCB, and applied our system to room-level place recognition. As a result, we confirmed that our system can accurately distinguish the eight places. language of the presentation: Japanese | |||
| 三﨑 慎也 | M, 2回目発表 | ユビキタスコンピューティングシステム | 安本 慶一, 岡田 実, 荒川 豊, 諏訪 博彦, 藤本 まなと |
|
title: Activity Recognition Method with Doppler Sensor Focusing on Frequency Characteristic and Motion Speed of Recognition Target
abstract: We propose an activity recognition method which uses motion sensors. In our previous work, we succeeded to recognize eight types of living activities by utilizing a power meter and location information. However, it has been difficult to accurately recognize activities that occur regardless of location such as “reading”, “using smartphone”, “using PC”, “eating” and “Stationary” regardless of location. One approach for recognizing these activities is to use the difference of motion speed of each activity as the feature quantity. In this presentation, We propose an effective method which can recognize the activities accurately by adding the motion speed of the target captured by the doppler sensor to the feature quantity. Also, we describe the evaluation experiments and future plans for the activity recognition method which uses doppler sensor. language of the presentation: Japanese 発表題目: 認識対象の動作速度と周波数特性に着目したドップラーセンサによる宅内行動認識手法 | |||