Year: 2025
Type: dissertation
Abstract: Dementia is a growing global health concern, particularly among the aging population. Traditional diagnostic methods for dementia rely heavily on clinical assessments, cognitive tests, and in-person observations, which can be resource-intensive, time-consuming, and often lead to late-stage diagnosis. The emergence of Internet of Things (IoT) technologies offers a promising solution by enabling continuous, real-time, and non-intrusive monitoring of elderly individuals in care facilities. This study explores the integration of IoT data with advanced visualization techniques to facilitate early detection and monitoring of dementia-related behaviours, thereby improving patient care and decision-making in elderly care environments. A prototype visualization dashboard was developed using Microsoft Power BI, leveraging IoT-based sensor data to track behavioural patterns, including sleep disruptions, activity transitions, and room occupancy trends. The system was designed to provide an intuitive, real-time interface for nursing staff and caregivers, offering interactive visual analytics, event playback functionalities, and anomaly detection features. Key features of the dashboard include: System monitoring: A real-time traffic light mechanism to indicate IoT sensor status and data integrity. Activity and dwell time analysis: Interactive visualizations that track movement patterns and detect deviations in patient behaviour. Sleep pattern insights: Line charts and heatmaps that highlight irregularities in sleep duration and continuity, which may indicate cognitive decline. Event mapping and playback: A dynamic process map and playback capability that reconstructs behavioural sequences for further analysis. To evaluate the effectiveness of the prototype, user testing was conducted with nursing staff in a simulated elderly care environment. The results demonstrated that the system improved caregivers' ability to monitor behavioural changes, detect anomalies, and make informed care decisions in a timely manner. The interactive and user-friendly design of the dashboard facilitated its adoption, ensuring that even non-technical users could interpret complex datasets efficiently. Furthermore, sleep pattern analytics proved particularly valuable in identifying early signs of dementia-related cognitive decline. Despite its promising outcomes, the study acknowledges certain limitations, including the reliance on passive infrared (PIR) motion sensors, which, while non-intrusive, do not capture qualitative aspects such as social interactions or facial expressions. Additionally, the sample size of end-users was limited, and future large-scale testing is required to assess its broader applicability in real-world care settings. Based on the findings, several avenues for future work are proposed, including: Integration with machine learning: Enhancing the system with AI-driven anomaly detection to provide predictive analytics for early dementia diagnosis. Expansion of IoT sensor networks: Incorporating additional wearable sensors (e.g., heart rate, temperature, gait tracking) to develop a more comprehensive health monitoring system. Decision Support System Development: Implementing automated recommendations based on behavioural patterns, allowing caregivers to intervene proactively. Scalability Testing: Conducting multi-resident deployment in elderly care facilities to assess long-term usability and clinical impact. In conclusion, this thesis work demonstrates the feasibility and potential of IoT-powered visualization tools in transforming dementia care. By bridging the gap between sensor-based monitoring and real-time decision-making, the developed dashboard paves the way for enhanced elderly care solutions. As the healthcare industry moves toward data-driven, AI-integrated solutions, IoT-based visualization tools like the one developed in this study hold great promise in improving patient outcomes, reducing healthcare costs, and optimizing caregiver workflows. Keywords: IoT in healthcare, dementia monitoring, visualization dashboard, elderly care, behaviour tracking, Power BI, real-time analytics, machine learning in dementia care, decision support system, IoT sensor data.
Authors: Islam, Muhammad Anwar Ul Haq, Kotty, Sri Venkata Naga Santhosh Ravi Kiran
Institution: Luleå University of Technology
Language: English
Subfield: Psychiatry and Mental health
Field: Medicine
Domain: Health Sciences
Keywords: dementia, early detection, elderly care, Internet of Things, sensor-based monitoring, visual analytics dashboard, Power BI, passive infrared motion sensor, cognitive decline, real-time monitoring, dwell time analysis, room occupancy monitoring, wearable sensors, decision support system, user testing
Open Access status: green
OpenAlex ID: W7025269317
