< Back to previous page


Health Behaviour Analytics On Heterogeneous Data

Several investigation have been carried out to monitor people’s activities at home. Most systems have been developed using either simple sensor data (wearable sensors, mobile phone accelerometer data) or camera information to recognise human activities in a home environment. However, existing work has either focus on simple domestic activities in real-life scenarios, or recognition of more complex activities in videos with well-defined situations. We still lack research on methods that can extract actionable insight for understanding relation between behaviour and lifestyle in the perspective of healthcare and providing personalized lifestyle coaching to improve health from multimodal and heterogeneous data. The goal of this dissertation is to address this scientific lock: developing a model of personal health pattern based on machine learning methodologies to the analysis of the multimodal data. We propose to control by reducing the scope of investigation to a specific situation, health-related behaviours measured by physiological sensors, in-home Internet-of-Thing (IoT) sensors, and devices. The Ph. D. project is organized around 4 main steps: • Recognising health-related activity based on multimodal data for the possible use case, with a specific focus on critical situation detection • Developing a tool for the efficient annotation of the multimodal data stream from multiple devices, including wearable sensors, IoT sensors and devices, and mobile phone sensors • Identifying the specific health behaviours and physiological parameters to extract relation between them for finding personal health pattern, through a multimodal data analysis. • Modelling the personal health profile, using machine learning techniques

Date:23 Sep 2018  →  Today
Keywords:machine learning, deep learning, heterogeneous data processing, activity recognition
Disciplines:Sensors, biosensors and smart sensors, Other electrical and electronic engineering
Project type:PhD project