Leveraging IoT and TinyML for Smart Battery Management in Electric Bicycles.

Ahstract-Urban mobility and environmental sustainability are critical challenges intensified by population growth, urbanization, and increased vehicle presence, leading to issues like congestion, pollution, and climate change. To counter this, the urgency for efficient and eco-friendly transportation solutions is evident. Battery Management Systems (BMS) play a central role, overseeing batteries in electronics, vehicles, and energy storage systems for safety and reliability. Focusing on electric bicycles (e-Bikes), the article highlights BMS's role in enhancing cyclist experiences and delves into techniques utilizing energy-efficient hardware and deep learning, notably TinyML, to predict Lithium-Ion battery State of Health (SoH). It examines diverse architectures and parameters, unveiling relationships to guide machine learning model construction, emphasizing delayed feedback integration based on rigorous statistics. The article navigates through data segregation and model training processes, presenting a comprehensive approach that balances sustainability, efficiency, and technological progress in predicting e-Bike battery SoH.