Lifecycle Extensions through Software-Defined Predictive Control of Power Electronics
Power electronics in electric vehicles are subject to various load fluctuations caused by the driver's behavior or by external conditions, such as the given altitude profile of the route traveled. The LETSCOPE project aims to directly integrate power electronics as part of an intelligent, predictive system management. On the one hand, the continuous recording of state variables for the state-of-Health of the power electronics, which will be reported to the intelligent management system, and on the other hand cloud-based services will be included to proactively influence the control of the power electronics by predicting the expected load and environmental changes. For implementation, a RISC-V SoC with AI accelerator will be developed, which will be used as a platform for both the classical control and the AI-based predictive part.
ITIV is designing a concept for the local computing and control system. This is to enable a connection to the sensors of the power electronics as well as the dynamic integration of cloud resources and other control devices. Predictive control including AI support is to be integrated into the system. Based on the specification, suitable AI algorithms will be selected and with a complexity estimation the partitioning between cloud and internal control units will be done.