A design platform for flexible programmable DSP for automotive sensor conditioning

In automotive field the signal processing algorithms implemented on chips is getting more and more complex. This is leading to a growing area of digital blocks of sensor systems and to greater digital design effort. The work presented in this paper has been started in order to have a flexible software-oriented DSP in portfolio, quickly customizable for the current product and with the minimum area overhead. The design platform has three levels: the instruction set emulator model for a fast algorithm implementation and the instruction set definition, the bit-true and cycle-true model of the processor for a fine verification of the application firmware, and the RTL design of a three-stage pipeline μ-processor. The instruction set and the hardware resources of the μ-processor are easily customizable. A case study, a design aided by the platform, is presented. Synthesis results are discussed, being the area constraint one of the most significant.