ENergy-Aware Sensing and DUty Cycling PRotocols for BattEryless Internet of Things using On-device and federaTed learning

Descrizione del progetto

Tecnologie avanzate di raccolta di energia per dispositivi IoT senza batteria

Le comunicazioni wireless di prossima generazione, incluso il 6G, sfrutteranno i dispositivi IoT che dovranno funzionare per anni consumando una quantità minima di energia. Tuttavia, le batterie attuali usano elementi delle terre rare che non possono essere prodotti e smaltiti in modo sostenibile. Nonostante i dispositivi IoT senza batteria offrano una soluzione potenzialmente più ecologica, trovare tecniche efficienti di raccolta dell’energia è ancora un ostacolo importante. Il progetto ENSURE-IoT, finanziato dal programma di azioni Marie Skłodowska-Curie, si propone di promuovere la sostenibilità nelle reti wireless sfruttando i progressi più recenti nelle tecnologie di recupero energetico e nelle innovazioni nell’ambito dell’IoT e dell’intelligenza artificiale.

Obiettivo

Future wireless technologies, such as sixth generation (6G) wireless networks, are relying significantly on low-power and low-cost Internet of Things (IoT) devices with extremely long lifetime. This increases the demand for batteries to power up IoT devices, which contain toxic substances that adversely affect the surrounding environment and ecological systems. The energy harvesting (EH) technology enables the development of batteryless IoT (BL-IoT) devices to promote environmental sustainability. However, the existing solutions for BL-IoT devices, such as sensing and communication protocols, do not account realistic operational scenarios and are bound to specific design parameters, which limit their performance and applicability in real-world scenarios. To address this, ENSURE-IoT aims to integrate EH technology with IoT and advanced Artificial Intelligence (AI) technologies to create novel, realistic, and intelligent energy-aware sensing and communication protocols based on theoretical models, simulation, and experimental validation for BL-IoT devices. Furthermore, the project will develop and deploy efficient and resource-friendly on-device learning and federated learning models to optimize decisions on both parts: the BL-IoT device to optimize decision accuracy locally, as well as on the access point under different circumstances. These AI models will facilitate BL-IoT devices to make energy-aware decisions in dynamic EH conditions. The performance of the developed solutions will be validated using theoretical modelling, simulation and hardware, and will be compared with benchmark protocols in dynamic EH conditions. The project results are expected to encourage wider adoption of BL-IoT solutions and to contribute to various Sustainable Development Goals, as well as EU policies, including the EU Green Deal, Net-Zero Industry Act, and Europe's Internet of Things policy.