The Jitter Margin and Its Application in the Design of Real-Time Control Systems

Abstract

Based on recent advances in control theory, we propose the notion of jitter margin for periodic control tasks. The jitter margin is defined as a function of the amount of constant delay in the control loop, and it describes how much additional time-varying delay can be tolerated before the loop goes unstable. Combined with scheduling theory, the jitter margin can be used to guarantee the stability and performance of the controller in the target system. It can also be used as a tool for assigning meaningful deadlines to control tasks. We discuss the need for best-case response-time analysis in this context, and propose a simple lower bound under EDF scheduling. Finally, a controlscheduling codesign procedure is given, where periods are assigned iteratively to yield the same relative performance degradation for each control task.