Effective battery systems are a cornerstone of the next wave of innovation. Emerging technologies including electric-drive vehicles (EVs) and stationary energy storage systems require integrated battery systems to provide sustained and reliable power. A unique opportunity exists for the software-controlled battery management system (BMS) to play a crucial role in enabling continued innovation. However, there are several unaddressed challenges to current BMS design. A battery's decreasing capacity over time due to aging necessitates that BMS automatically adapt to battery changes. Furthermore, future BMS systems will require autonomous reasoning capabilities to make economically-sound decisions on users' behalf (e.g., scheduling battery charging times in a personalized fashion). This project injects intelligence capabilities into BMS design with the development of the Autonomous Battery Operating System (ABOS). ABOS enables more energy-efficient, long-lasting, and secure battery-driven systems. Furthermore, the PIs incorporate computational and cyber-security aspects of ABOS design into their undergraduate- and graduate-level courses.
ABOS advances the science of autonomous system design with simultaneously introspective and extrospective learning. ABOS will learn and adapt to user-initiated charging/discharging patterns, reason about how these patterns affect the battery's state-of-health, and respond to potential faults or attacks. The PIs and their external collaborators will develop a hybrid EV simulation environment to test the ability of ABOS to control a physical battery system. The simulation environment will evaluate the effectiveness of ABOS in predicting battery state, minimizing cost of operation, and handling failures and threats.
