The research directions below aim to advance human-centred intelligent home environments that are adaptive, trustworthy, sustainable, and supportive of independent living and ageing in place. By integrating insights from computer vision, human–computer interaction, and embedded AI, my work seeks to design systems that not only understand and respond to human activity but also respect privacy, support user autonomy, and promote sustainable living. The following are my key research directions:
Human Activity Analysis in Home Environment
This research focuses on developing novel deep learning models for analysing human activity within home environments. The aim is to address both the technical and human-centered challenges that influence user acceptance and system performance in home activity monitoring. Key challenges include balancing the trade-off between accuracy and computational efficiency to enable real-time monitoring on low-cost, resource-constrained embedded systems; ensuring privacy and data security; and supporting user agency and autonomy in how monitoring systems operate within personal spaces.
User Acceptance of Ambient Assisted Living Technologies
This research explores the human factors that shape the adoption and sustained use of ambient assisted living (AAL) technologies in home environments. It seeks to understand user perceptions, preferences, and concerns regarding these systems, with a focus on identifying the barriers and facilitators to their acceptance. The overall goal is to ensure that AAL technologies are designed to meet the needs, expectations, and ethical considerations of end users, thereby promoting trust, adoption, and long-term engagement.
Computer Vision for Smart & Sustainable Homes
This research investigates computer vision–based methods for occupant-centred control of home systems to achieve a balance between thermal comfort and energy efficiency. It focuses on developing non-invasive, privacy-preserving approaches to capture and interpret occupant data. Deep learning models are trained to estimate comfort levels, which are then used to automate systems such as HVAC in a way that enhances both user comfort and sustainability.