Robust Controller Scheduling in Automotive Communication Networks

The increasing demand in the automotive industry for more services, security, economy and comfort is characterized by an increasing number of Electronic Control Units (ECUs) utilized and an increasing complexity on the communication between them. In such distributed control systems, sensors, actuators and controllers communicate through a shared medium(the communication bus) which inevitably constrains the amount of bandwidth available. In this paper, we consider the robust off-line (static) communication schedule design problem. Given a (distributed) plant and a controller, we design a (sub)optimal communication policy between `nodes' that minimizes the performance degradation introduced by the network, using an H-infinity framework. Uncertainty and nonlinearities in the plant can be easily catered for by the approach. The complex optimization process is performed by fast versatile algorithms that can guarantee schedule calculation close to the optimum and that can easily handle scheduling constraints such as bandwidth demands on selected ECUs. Theoretical results are validated through numerical analysis and practical implementation in a hardware-in-the- loop vehicle control system which can be generalized to a FlexRay or TTCAN environment.