The systematic design of automotive control applications is a challenging problem due to lack of understanding of the complex and tight interactions that often manifest during the integration of components from the control design phase with the components from software generation and deployment on actual platform/network. In order to address this challenge, we present a systematic methodology and a toolchain using well-defined models to integrate components from various design phases with specific emphasis on restricting the complex interactions that manifest during integration such as timing, deployment, and quantization. We present an experimental platform for the evaluation and testing of the design process. The approach is applied to the development of an adaptive cruise control, and we present experimental results that demonstrate the efficacy of the approach.

Designing cyber-physical systems (CPS) is challenging due to the tight interactions between software, network/platform, and physical components. A co-simulation method is valuable to enable early system evaluation. In this paper, a cosimulation framework that considers interacting CPS components for design of time-triggered (TT) CPS is proposed. Virtual prototyping of CPS is the core of the proposed framework. A network/platform model in SystemC forms the backbone of the virtual prototyping, which bridges control software and physical environment. The network/platform model consists of processing elements abstracted by realtime operating systems, communication systems, sensors, and actuators. The framework is also integrated with a model-based design tool to enable rapid prototyping. The framework is validated by comparing simulation results with the results from a hardware-in-the-loop automotive simulator.

No abstract