Knowledge Distillation and Federated Learning for Data-Driven Monitoring of Electrical Vehicle Li-Battery

In last recent years, deep neural networks have seen success in Electric Vehicles (EVs) monitoring. Their scalability with large-scale data and numerous model parameters contributes to this success. However, EVs use resource-constrained devices that struggle with complex models and data privacy concerns prevent sharing outside the owning device. Thus, Knowledge Distillation (KD) and Federated Learning (FL) emerged as solutions for models simplification and for facilitating distributed training on private data. We propose and compare architectural solutions leveraging KD and FL to address challenges in EV environments, where resources are limited, and data privacy is crucial. By simplifying models with KD and enabling distributed training with FL, our architectures aim to integrate Machine Learning (ML) models into Internet of Things (IoT) devices within EVs, enhancing monitoring and predictive maintenance applications.