AbstractWe present a test quality measure that allows for quantifying the completeness of black-box tests for continuous-time dynamic systems. The measure is based on a state space model of the system under test. The metric has been called the state space coverage. The classical coverage metrics, such as statement, branch, and path coverage, are not appropriate for dynamic systems because such systems are defined by differential equations and usually have an infinite number of states. The objective of the paper is to develop a necessary foundation for the metric as well as to present guidance on its application to software systems that incorporate dynamic behavior. The purpose of the proposed solution is to better assure the test engineer that a given test set is sufficient and to indicate where additional testing is required. An application example is presented to illustrate theoretical analysis and mathematical formulation.
