Extending IoT Devices Network Lifetime Using IoT Service Prediction and Recommendation

IoT devices continue to be an integral part of our everyday lives, providing convenience, efficiency, and enhanced capabilities. With limited storage, computational, and energy resources, the optimized management of these resources becomes essential for resource-constrained devices. The proposed solution is a two-phase framework for optimizing IoT resource management. First, we apply a Time-Series analysis Machine Learning (ML) technique to predict upcoming user service requests based on their historical interactions and preferences within a specified Region of Interest (RoI). Second, using anticipated user requests, we design a recommendation system to suggest a set of devices that serve those service requests based on the user’s Quality of Service (QoS) and security preferences, while addressing common challenges such as the cold-start problem, device availability, and energy consumption. The validity of our framework is characterized by an improved device network lifetime, with devices being available up to 95% of the service time, a 20% increase in the number of user requests fulfilled, and the ability to provide more service with fewer devices. These results highlight that predicting future user service requests, along with a service recommendation method, is a novel way to optimize the resource management process.