JITTERBUG: A tool for analysis of real-time control performance

The paper presents JITTERBUG, a MATLAB-based tool-box for real-time control performance analysis. The control system is described using a number of con-nected continuous-time and discrete-time linear sys-tems driven by white noise. The control performance is measured by a continuous-time quadratic cost function. A stochastic timing model is used to de-scribe when the different discrete-time systems are updated during the control period. Building different models, the tool makes it easy to investigate how the control performance is affected by e.g. delay, jitter, lost samples, aborted computations, and jitter com-pensation. Aperiodic and multi-rate controllers may also be studied. The tool is also capable of computing the spectral densities of the different signals in the system. 1