CRII: CPS: Minimizing the Oracle Problem for Self-Adaptive Cyber-Physical Systems

Oakland University, Rochester MI

Investigators

Abstract

This project investigates how to ensure that self-adaptive cyber-physical systems (SA-CPS) function correctly as they execute. An SA-CPS can reconfigure itself, at run time, to mitigate unforeseen circumstances, where an example SA-CPS is an autonomous vehicle that reacts to unexpected road conditions by automatically adjusting its driving strategy to avoid an icy patch. Therefore, the system must be tested to ensure that it operates correctly following a reconfiguration. This project will provide techniques for ensuring that verification and validation strategies for an SA-CPS have been derived correctly and that any future adaptations are correct with respect to its original requirements. This research will enhance assurance for SA-CPSs by providing test oracles (i.e., software that determines if a test passes or fails) that adapt alongside an executing system; verifying and validating adaptive test oracles at design time and run time; and providing traceability between adaptive oracles and software requirements. Together, these three points will ensure that both test cases and oracles can effectively validate a running system throughout its lifetime. This work will be demonstrated on real hardware to show that together, software verification and validation can enhance system assurance as well as public perception and trust in SA-CPSs. This project will ensure that software test oracles are verified and validated to ensure that any software validation activities are being correctly performed on SA-CPS. The knowledge gained from this project will be published in highly-regarded conferences/journals and all artifacts and results released to GitHub and NSF's CPS Virtual Organization (CPS-VO). Information will also be shared publicly through a series of blog posts intended to educate the public on CPSs. This project will also train graduate students, results will be presented in a graduate class on CPS, and distilled into summer camps to introduce K-12 students to embedded programming.