Hardware/software co-simulation for the rapid prototyping of an acceleration sensor system with force-feedback control

We present hardware-in-the-loop simulations of a MEMS acceleration sensor system with view to reduce the development time and complexity especially regarding design and test of dedicated control algorithms. To this end, the physically-based high-level Simulink model of the mechanical sensor is connected to a FPGA hardware board, on which the corresponding control algorithms (in our case PID and fuzzy logic) are implemented. The different control algorithms are described on the level of behavioral modeling in Simulink using the Xilinx System Generator tool, which then automatically generates a VHDL code of the models and transfers it to the hardware board. The advantage of these hardware/software co-simulations compared to "pure" simulations of the entire system in Simulink is a realistic estimation of the overall performance (run times, power consumption, efficiency, etc.) of the control circuitry even before the transducer is realized in hardware itself. This, on the other side, allows also modifications and optimizations not only on the control circuit but also on the transducer design in a very early stage of the design process.