Constraint-driven synthesis and tool-support for FlexRay-based automotive control systems

Emerging bus protocols such as FlexRay provide an expedient platform for the design of automotive control systems due to its high bandwidth and deterministic temporal behavior. However, the choice of suitable platform parameters such as task and message schedules becomes a challenging design problem as the protocol is complex in nature and enforces a tight coupling between local task schedules on ECUs and global bus schedules. Although there exist several commercial off-the-shelf (COTS) design tools for FlexRay and control systems, current tools do not provide any mechanism for automatically synthesizing the platform parameters from the controller specifications. In this work we synthesize controllers subject to specified control goals while taking into account platform-specific properties. In particular, we translate the timing constraints derived from the control design into platform constraints that need to be satisfied by the control-related tasks and messages. For this purpose, we formulate and solve a constraint satisfaction problem (CSP) to synthesize feasible platform parameters that can be realized by the underlying operating systems and the FlexRay bus. Our design flow may be easily integrated with existing FlexRay design tools and will significantly ease (and automate) the existing design process. We show the applicability of our results by implementing two automotive control systems on a Hardware-in-the-Loop (HiL) setup and study how different bus configurations affect the controller synthesis and the choice of platform parameters.