Scaling Compute-Intensive Telehealth Monitoring in IoT-Enabled Smart Spaces: A Personal-Space-State Approach

IoT has created smart spaces that integrate sensory data, processes, objects, and people to enable continuous monitoring that provides knowledge for diverse services. These optimized spaces are essential for many safety and mission-critical monitoring applications, e.g. assisted living and healthcare. In this paper, we address a key challenge in monitoring smartspace at scale, i.e. that of energy expenditure associated with continuous compute-intensive operations. Diverse individuals need monitoring applications at different levels of quality based on their needs; we exploit these quality/needs tolerances by modeling them as a “personal-space-state”, a unique abstraction for spaces/individuals contexts. Then, we intelligently leverage the dynamic personal-space-state to provision services (multiple algorithms with different configurations) to reduce energy overhead while ensuring application quality. We propose an efficient IoT strategy that triggers actions and personal-space-state shifts to drive energy-optimized services for each individual. To validate our approach, we derive use-cases from real-world assisted living facilities with multiple monitoring applications such as fall detection and gait monitoring. Through detailed testbed measurements and larger simulated scenarios, we show that our personal-space-state transition approach can achieve greater than 2X reductions in energy dissipation and computational overhead without loss of accuracy.