Research Challenges

Internet of Things (IoT) enabled Healthcare is defined as a specific research area focusing on the utilization of IoT enabled techniques to offer high quality health services, including faster and safer preventive care, lower overall cost, improved patient-centred practice and enhanced sustainability. Current work on IoT enabled healthcare is highly interdisciplinary involving methodologies from computing, engineering, information science, behaviour science, as well as many different areas in medicine and public health. A promising trend in these studies appears to be developing sophisticated techniques that will enable:

• Integrating different smart wearable devices into a unified system for intelligently sensing daily human health information.

• The design and development of IoT enabled healthcare system or applications for efficiently delivering specific health services.

• Effectively and efficiently managing, analysing and exploring a sheer volume of long-term health data for supporting wise clinical decision-making.

However, addressing this trend is still significantly challenging due to an array of factors that include: shortage of cost-effective and accurate smart medical sensors, unstandardized IoT system architectures, heterogeneity of wearable devices connected, multi-dimensionality and high volume of data generated, and high demand for interoperability. Additionally, successfully empowering the utility of IoT enabled technology in healthcare will need an interoperable IoT environment for care delivery and research, tightly-coupled health data mining applications, adequate data and knowledge standards of self- empowerment and sound clinical decision-making foundation. These challenges and needs require the design and development of a series of innovative and comprehensive informatics methods in IoT enabled healthcare. 

Current research

Currently, I lead the Pervasive Computing Group consisting of 6 academics, 7 PostDoc and 15 PhD students working on this area. Main research goal was to investigate and develop an efficient data collection utilities solution supporting heterogeneous wearable sensors.

In MyHealthAvatar project, it involves developing data collection utilities for support data contribution from users with minimal input, building a data repository to store health related data of individual citizens collected from the web and mobile apps, and experimenting the linked data approach to facilitate the flourish and reuse of data, including data search and reasoning.

In MyLifeHub project, it involves research works on investigating new techniques enabling simultaneously and long-term quantifying the functional impairment related to vision underpinned with smart glasses; designing a platform to assess the impact of visual impairment on the QoL of ophthalmic patients both in general health terms and in vision specific terms; studying and proposing the data mining techniques for analysing and exploring the lifelogging data.

We are currently working in two directions below:

• how to use advanced machine learning approaches to effectively match uncertain life-logging data to high level personal life pattern and physical activities.

• investigate IoT-enabled technologies for providing a cost-effective and non-intrusive intelligent system to monitor and analyse dementia related behaviour in a home setting. 

Our future work will primarily focus on 3 aspects:

• 1. How to analysis and explore these IoT based health data with advanced algorithms. This problem will be closely related to big data analytics and data science. One research problem is how to use advanced machine learning approaches to effectively match uncertain life-logging data to high level personal life pattern. This problem involves the investigation of data mining approaches on missing and noisy data. General machine learning methods like decision trees, random forest, and neural networks can be evaluated in this issue. It also needs to use text mining methods for social media analysis.

  2. How to integrate structured and unstructured data collected from diverse sources and used as an ensemble to derive information. In IoT enabled platform, there are enormous real-time data streams from various sources, located in different servers. All data streams may be considered as ephemeral but may be in unknown servers. The following areas would be interested: Semantic approaches to support machine processing of data, big data knowledge extraction through ontology mapping, integrating data across vocabularies, mining data to discover rules about the data that can then be implemented in computational systems. Also, it needs to investigate some existing and novel knowledge representation and reasoning methods.

  3. How to effectively and efficiently linking and analysing long-term personal health data with hospital EHR data for supporting wise clinical decision-making. Normally the personal health data is significantly different with hospital HER data, regarding the types, content, frequency of collection, etc. At the moment, IoT enabled healthcare platforms suffer with difficulties of security and trust to connect and share data with hospital healthcare system. So, it needs to establish novel methodologies for reasoning and management of medical knowledge and use the methodologies to create and evaluate measures of patient long-term personal profile similarity based on mined temporal patterns in longitudinal patient records.