Design and verification of Cyber-Physical Systems using TrueTime, evolutionary optimization and UPPAAL

摘要

Timing imperfections in Cyber-Physical Systems (CPS) components affect their performance and reliability. This investigation presents a methodology to design and verify CPS using multi-objective evolutionary optimization, model checking and supporting software tools. The time-varying delays in CPS are modeled as constant delays plus jitter. It is shown that the CPS design problem is a trade-off between performance and jitter margin. To make good trade-offs, this investigation models the design problem as a constrained multi-objective optimization problem. The design algorithm results in a non-linear and multi-objective optimization problem that is generally difficult to solve. To overcome this computation barrier, this investigation uses the evolutionary algorithm Non-dominated Sorting Genetic Algorithm II (NSGA II). Solution of the optimization problem computes the proportional integral and derivative controller gains that simultaneously maximize the jitter margin while improving the system performance. Implementing the CPS controller in embedded platform requires the selection and validation of the process schedules as well as network protocols. To validate the CPS design, the computed controller gains are simulated in TrueTime to verify the performance for a given network protocol and processor scheduling policy. Three scheduling policies are considered due to their appropriateness for use with temporal imperfections: fixed priority, earliest deadline first and deadline monotonic. Finally, to verify the CPS for timing guarantees, a timed-automata model is used that defines the timing interfaces among the components. The formal model is then used to verify CPS response time and properties such as safety using computation tree logic in UPPAAL model checker. The proposed CPS design and verification approach is illustrated on an industrial mine pump example. Our results demonstrate that the proposed approach can be used to design and validate CPS for performance and verify timing guarantees. The proposed method provides a systematic design and verification approach that can be used for deployments of CPS in industries. (C) 2016 Elsevier B.V. All rights reserved.