Automatic Code Generation - Technology Adoption Lessons Learned from Commercial Vehicle Case Studies

Abstract

Using Model-Based Design, engineers model complex systems and simulate them on their desktop environment for analysis and design purposes. ModelBased Design supports a wide variety of C/C++ code generation applications that include stand-alone simulation, rapid control prototyping, hardware-in-theloop testing, and production or embedded code deployment. Many of these code generation scenarios impose different requirements on the generated code. Standalone simulations usually need to run fast, for parameter sweep or Monte Carlo studies, but do not need to execute in true hard real-time. Hardware-in-the-loop tests by definition use engine control unit (ECU) component hardware that requires a hard real-time execution environment to protect the physical devices. Code generated for production ECUs must satisfy hard real-time, efficiency, legacy code, and other requirements involving verification and validation efforts. With Model-Based Design, the functional behavior of the model needs to match that of the generated code. As a result the transformation of models into generated code must include necessary deployment and real-time artifacts to ensure that the code executes properly in the final software and hardware environments. For example, in a typical commercial vehicle use case, a diesel engine control algorithm and engine plant model are simulated together as a hybrid system. The plant model is input into the code generator for deployment in a hard real-time HIL lab. Code generation for the engine control algorithm is often done in two, or even three, phases. First, the code is generated for real-time rapid control prototyping for algorithm assessment and refinement. Next, the code may be generated for execution on the actual embedded microprocessor during on-target rapid prototyping for algorithm assessment on the ECU hardware. Finally, the code is generated for production ECUs and several verification steps are employed, including software-in-the-loop (SIL), processor-in-the-loop (PIL), and finally hardware-in-theloop (HIL) testing. Organizations moving from traditional waterfall processes that involve paper documents and hand code to Model-Based Design face challenges familiar to those who have followed other technology migrations, such as drafting tables to CAD systems or Assembly language to C code. These challenges center on how to: • best leverage the technology • reuse existing process • pace the transition • develop necessary skills sets and training This paper describes case studies on how John Deere adopted Model-Based Design for commercial vehicle development and discusses the benefits and lessons learned. INTRODUCTION TO MODEL-BASED DESIGN A model represents a dynamic system whose response at any time is a mathematical function based on its inputs, current state, and current time. Historically, system engineers have used block diagrams as shown in Figure 1 to create models and design algorithms within numerous engineering areas such as Feedback Control and Signal Processing. In recent years, graphical modeling environments consisting of block diagrams and state machines have been used to analyze, simulate, prototype, specify, and deploy software algorithms within a variety embedded systems and applications. ModelBased Design refers to the use of models and modeling environments as the basis for embedded system development. Output Plant Environment Embedded System Input