AI Virtual Sensors: A Revolution in Battery Management

A recent webinar highlighted the innovative use of virtual sensors based on artificial intelligence (AI) to estimate signals that are often difficult or costly to measure. A concrete example is the state of charge (SOC) of batteries, which is crucial in battery management systems. This session demonstrated how AI models can be integrated into system-level design and validated while considering performance, available resources, and deployment constraints.

The process detailed in the webinar shows how to design, verify, compress, and deploy these AI virtual sensors on embedded processors, all within a unified environment. Participants learned how to integrate AI models into Simulink® to simulate and verify systems at a global level.

Among the skills taught, formal verification plays a key role in assessing the behavior of neural networks. Additionally, model optimization is essential for reducing memory footprint and improving execution speed, which is critical for embedded deployment. The webinar also covered code generation and profiling without libraries, thus facilitating deployment on embedded systems.

Finally, participants were able to evaluate design trade-offs between accuracy, performance, and deployment objectives, an essential aspect for maximizing the efficiency of battery management systems. To access this content, it is necessary to create a user profile on the dedicated hub and log in to enjoy all available resources.