A Cloud-Integrated Predictive Control Evaluation System
In the LETSCOPE project, project partners investigate ways to enhance the lifetime of power semiconductors in the drive system of electric cars. The envisioned solution will use AI to combine cloud data and local data sources to realize predictive control, i.e. knowing ahead of time what kind of load change will occur in the system. Prof. Becker’s group at ITIV provides an FPGA-based prototyping platform to enable both AI acceleration, interfacing local sensors and cloud as well as execution of the traditional control algorithms in an real-time system. The prototyping FPGA-based system shall be realized with the possibility to transfer the design to an ASIC chip later on.
The task of this thesis is to design, implement and evaluate a hardware-in-the-loop (HIL) test and evaluation system for the LETSCOPE FPGA prototype. First, a literature research on HIL systems shall be carried out and interfaces to be emulated (CAN, WIFI, …) shall be selected. CAN and WIFI extension boards shall be selected. Drivers for the devices shall be developed and integrated. Then a HIL simulation system running on the PS shall be designed and implemented. The HIL system shall emulate CAN messages and cloud services and capture the system’s control outputs. To evaluate the HIL system, some tests and demonstrations of the predictive control system shall be done.
- Embedded C++ knowledge recommended
- Python (GUI) knowledge recommended