A framework for the co-simulation of engine controls and task scheduling

Abstract

To evaluate the impact of scheduling latency and task design on the performance of engine control applications, we developed a co-simulation framework, based on Simulink and an extension of the T-Res scheduling simulator tool. The objective of the research and the tool development is to provide a better characterization of the very popular problem of scheduling and analysis of Adaptive Variable Rate Tasks (AVR) in engine control. The purpose of the tool is to go beyond the simplistic model that assumes hard deadlines for all tasks and to study the impact of scheduling decisions (and possibly missed deadlines) with respect to the functional implementations of the control algorithms and the engine performance. The developments include a co-simulation framework and a set of models for the engine components in order to evaluate the performance with respect to fuel efficiency, consumption, soot and NOx emissions.