CPS Engineering: Gap Analysis and Perspectives

TL;DR

This paper analyzes the challenges in verifying and certifying complex cyber-physical systems (CPS) and CPSoS, emphasizing control methods and proposing research directions for scalable verification strategies.

Contribution

It identifies key open verification problems in CPS/CPSoS and reviews promising research threads for scalable, long-term verification and certification approaches.

Findings

Set-based and invariant-based design methods show promise.

Contracts and adversarial testing are effective for safety assurance.

Probabilistic estimation aids in scalable verification.

Abstract

Virtualization of computing and networking, IT-OT convergence, cybersecurity and AI-based enhancement of autonomy are significantly increasing the complexity of CPS and CPSoS. New challenges have emerged to demonstrate that these systems are safe and secure. We emphasize the role of control and emerging fields therein, like symbolic control or set-based fault-tolerant and decentralized control, to address safety. We have chosen three open verification problems we deem central in cost-effective development and certification of safety critical CPSoS. We review some promising threads of research that could lead in the long term to a scalable and powerful verification strategy. Its main components are set-based and invariant-based design, contracts, adversarial testing, algorithmic geometry of dynamics, and probabilistic estimation derived from compositional massive testing. To explore these orientations in collaborative projects, and to promote them in certification arenas, we propose to continue and upgrade an open innovation drone-based use case that originated from a collaborative research project in aeronautic certification reformation