상세정보
학술대회
HaaS(Human Activity Analytics as a Service) Using Sensor Data of Smart Devices
- 발행일
- 201810
- 출처
- International Conference on Information and Communication Technology Convergence (ICTC) 2018, pp.1500-1502
- 협약과제
- 18HR1800, 다중로그 기반 멀티모달 데이터융합 분석 및 상황 대응 플랫폼 기술 개발, 이용태
- 초록
- recognizing human activities with sensor data of smart device is challenging task due to issues, such as data privacy, lack of dataset size, and appropriate preprocessing techniques to eliminating noise of time-series data. In this paper, we present a human activity recognition analytical models using time-series formed sensor data of smart devices such as smartphone and wearable devices. In order to understand user's activities within an environment existing various devices, it needs to collect data generated in that specific environment and analyze it with guaranteeing high performance and accuracy for classifier.This paper makes two specific contributions: first, we present a novel system architecture for recognizing human activities. This architecture comprises of a data collection protocol, which carries sensor data of smartphones and wearable devices into a server platform; second, we show experiment results comparing with three kinds of analytical models and describe its meaning.
- KSP 제안 키워드
- Analytical model, Data Collection, Experiment results, High performance, Human activity recognition(HAR), Preprocessing techniques, Smart devices, System architecture, Time series data, Wearable device, data privacy