دانلود رایگان مقاله طبقه بندی و مناسب بودن فناوری سنجش برای به رسمیت شناختن فعالیت

Abstract

Wider availability of sensors and sensing systems has pushed research in the direction of automatic activity recognition (AR) either for medical or other personal benefits e.g. wellness or fitness monitoring. Researchers apply different AR techniques/algorithms and use a wide range of sensors to discover home activities. However, it seems that the AR algorithms are purely technology-driven rather than informing studies on the type and quality of input required. There is an expectation to over-instrument the environment or the subjects and then develop AR algorithms, where instead the problem should be approached from a different angle i.e. what sensors (type, quality and quantity) a given algorithm requires to infer particular activities with a certain confidence? This paper introduces the concept of activity recognition, its taxonomy and familiarises the reader with sub-classes of sensor-based AR. Furthermore, it presents an overview of existing health services Telecare and Telehealth solutions, and introduces the hierarchical taxonomy of human behaviour analysis tasks. This work is a result of a systematic literature review and it presents the reader with a comprehensive set of home-based activities of daily living (ADL) and sensors proven to recognise these activities. Apart from reviewing usefulness of various sensing technologies for home-based AR algorithms, it highlights the problem of technology-driven cycle of development in this area.

4. Discussion

Is it possible for one sensor technology to discover all of the home activities with 100% accuracy? Can this sensor technology be used in both single-occupancy and multi-occupancy homes? The results of the systematic literature review presented in this paper show that an affirmative answer cannot be given for any of the reviewed sensor technologies when used on their own. Video sensors and accelerometers have the highest potential but with the current state-of-the-art they are not capable of recognising the full range of activities without even considering the practicalities of using these technologies, as they are not suited for every part of the house. Cameras would not be acceptable in areas such as bathrooms and bedrooms; for practical reasons accelerometers cannot be given to every visitor who is not living in the house. The conclusion is that currently no single sensor/sensing technology can discover all home-based ADLs. Accelerometer-based wearable sensors are promising, yet need some contextual information to differentiate between activities such as preparing tea and coffee. The solution to the problem lies in multi-modal IoT sensor systems that take into account basic principles of ubiquitous computing. It is important to design smart-spaces in such a way as to avoid overinstrumentation of both the space and subjects with redundant sensors. Fundamentally, whilst in this paper we focused only on IoT data collected in home environments, the question remains on how to establish the value of the IoT data (and, therefore, the accepted value of the IoT infrastructure required to acquire and make this data available) even before the data are used to infer some information. The SPHERE [126] project is addressing this question by carrying out quantitative evaluation of which sources of data provide best impact, according to defined metrics, on known AR algorithms.