Resource-aware Automotive Control Systems Design: A Cyber-Physical Systems Approach

Abstract

As the automotive industry is entering the smart era through advancesin sensing, computation, storage, communication, and actuation technologies,a larger number of more complex control applications withbetter performances are expected to be on board. This requires an implementationplatform with abundant resources, which is undesired inthe cost-sensitive automotive domain. The implementation platform,often embedded in an Electronic Control Unit ECU and shared bymultiple applications to save cost, is mainly comprised of a processorfor computation, memory for storing instructions and data, and busfor internal and external communication. Conventionally, automotivecontrol systems are designed using model-based approaches, where thedetails of the implementation platform are ignored. Techniques thatintegrate the characteristics of implementation resources into controlalgorithms design are largely missing. Such a separate design paradigmis too conservative in resources dimensioning and utilization for modernvehicles. This article presents recently developed approaches in automotivecontrol systems design that take implementation resources intoconsideration, aiming to improve the control performances for a givenamount of resources, or equivalently, realize the required control performanceswith fewer resources. While communication resources have beenextensively explored in the literature of networked embedded controlsystems, we will focus on memory and computation resources, whichhave started to receive attention from the academic community andindustry just recently. As Electric Vehicles EVs have become a newtrend in the automotive industry, energy resources of EVs, i.e., thebatteries, are also investigated. A number of real-world applicationsvalidate the resource-aware automotive systems design techniques presentedin this article.