Automated Integration of MathWorks® Simulink® Signal Flow Graph Models into Synopsys® VirtualizerTM-based Virtual Prototypes

Authors: Andreas Mauderer, Robert Bosch GmbH, DE; Jan-Hendrik Oetjens, Robert Bosch GmbH, DE; Alexander Schreiber, The MathWorks GmbH; Jerome Chevalier, The MathWorks, Inc.

„2nd Place - Best Paper“-Award


The rapid progress of Advanced Driver Assistance Systems (ADAS) increases the importance of automotive sensors, leading to stringent requirements and reduced time-to-market. MEMS sensors are usually integrated as System-in-Package (SiP), with a signal processing System-on-Chip (SoC) ASIC (Application-Specific Integrated Circuit). Virtual Prototyping is an established approach for early validation of complex SiP/SoC concepts and for early software development. In this paper, we present an improvement to our IEEE 1685™: IP‑XACT‑centric workflow for Virtual Prototype (VP) creation [1]. This approach was developed in close collaboration between Bosch and MathWorks®. In the approach, extended IP‑XACT register descriptions provide input to MathWorks code generation tools to automatically produce C/C++ code for e.g. signal processing functionality and SystemC® TLM 2.0 wrappers with Synopsys® SystemC Modeling Library (SCML) registers. These components, designed and verified using Model-Based Design (MBD) and MathWorks MATLAB® and Simulink® software, can directly be integrated into a Synopsys® Virtualizer™ architecture. We show how the proposed approach accelerates VP construction time and ensures functional consistency by applying it to an industrial-grade ASIC design.

Publication Date: 2016/06/23

Location of Publication: Synopsys Users Group, SNUG Germany 2016, Hilton Munich City, Germany

Keywords: Automotive; Verification