Testing real-time systems under uncertainty

Model-based testing is a promising technique for improving the quality of testing by automatically generating an efficient set of provably valid test cases from a system model. Testing embedded real-time systems is challenging because it must deal with timing, concurrency, processing and computation of complex mixed discrete and continuous signals, and limited observation and control. Whilst several techniques and tools have been proposed, few deals systematically with models capturing the indeterminacy resulting from concurrency, timing and limited observability and controllability. This paper proposes a number of model-based test generation principles and techniques that aim at efficient testing of timed systems under uncertainty.