Advanced state of charge evaluation techniques for embedded battery management systems

Lithium-ion batteries are becoming increasingly popular due to their excellent characteristics. However, to reduce safety risks and avoid failures, the State of Charge (SoC) must be accurately monitored by the Battery Management System (BMS). Unfortunately, the SoC indicator cannot be measured but has to be properly estimated by suitable algorithms. Several techniques are available in the literature, showing advantages and disadvantages depending on the application field and the required accuracy. In this dissertation, some advanced techniques, including hybrid approaches and Machine Learning, have been investigated. The aim was to implement the algorithms on an embedded system and evaluate their effectiveness in terms of computational resources and SoC estimation accuracy in a real environment.