Stochastic Model-based Analysis of Timing Errors for Mechatronic Systems with User-defined General Discrete-time Distributions

Abstract

This paper introduces a new stochastic method for model-based analysis of the timing behavior for reliable design of mechatronic systems with distributed, concurrent processes. Given a baseline behavioral system model, e.g. a semi-formal UML activity diagram, which actions are annotated with timing properties specified with user-defined general discrete-time distributions and formalized timing requirements, this method helps to analyze the occurrence of timing errors. The method comprises a specific baseline model reduction, the mapping of the semi-formal model into a formal stochastic Petri net model, and the generation of a discrete-time Markov chain model. The results allow the design verification and the comparison of various design options regarding the timing behavior in early design phases. A concept study of a mobile medical patient table serves as a demonstrative example.