CDS: "Jitterbug and TrueTime: Timing analysis and simulation of networked embedded control systems"

Karl-Erik Arzen
Professor at the Department of Automatic Control
Lund University, Sweden

The temporal non-determinism introduced by computing and communication in the form of delay and jitter can lead to significant performance degradation. To achieve good performance in systems with limited computer resources, e.g.embedded systems, the resource constraints of the implementation platform must be taken into account at design time. To facilitate this, software tools are needed to analyze and simulate how the timing affects the control performance. Two such tools, JITTERBUG and TRUETIME have been developed at the Department of Automatic Control, Lund University over the last five years.

JITTERBUG is a MATLAB-based toolbox that makes it possible to compute
a quadratic performance criterion for a linear control system under various timing conditions. The tool can also compute the spectral density of the signals in the system. Using the toolbox, one can easily and quickly assert how sensitive a control system is to delay, jitter, lost samples, etc., without resorting to simulation. The tool is quite general and can also be used to investigate jitter-compensating controllers, aperiodic controllers, and multi-rate controllers. The main contribution of the toolbox, which is built on well-known theory (Markov jump linear systems), is to make it easy to apply this type of stochastic analysis to a wide range of problems.

The use of JITTERBUG assumes knowledge of sampling period and latency
distributions. This information can be difficult to obtain without access to measurements from the true target system under implementation.Also, the analysis cannot capture all the details and nonlinearities (especially in the real-time scheduling) of the computer system. A natural approach is to use simulation instead.  The MATLAB/Simulink- based tool TRUETIME facilitates simulation of the temporal behaviour of a multitasking real-time kernel executing controller tasks. The tasks are controlling processes that are modelled as ordinary Simulink blocks.  TRUETIME also makes it possible to simulate simple models of wired and wireless communication networks and their influence on networked control loops. Different scheduling policies may be used (e.g., priority-based preemptive scheduling and earliest-deadline-first (EDF) scheduling).

During the presentation the tools will be demonstrated.