ROFLSEJul 7, 2020

Offline Runtime Verification of Safety Requirements using CSP

arXiv:2007.03522v32 citations
Originality Synthesis-oriented
AI Analysis

This work addresses safety verification for industrial systems like fusion reactors, but it is incremental as it builds on existing runtime verification and CSP modeling techniques.

The paper tackles the challenge of ensuring ongoing safety compliance in systems by presenting Varanus, a runtime verification toolchain that uses a CSP model derived from safety requirements, and demonstrates its application on the MASCOT teleoperated manipulation system, revealing an underspecification in the design.

Dynamic formal verification is a key tool for providing ongoing confidence that a system is meeting its requirements while in use, especially when paired with static formal verification before the system is in use. This paper presents a workflow and Runtime Verification (RV) toolchain, Varanus, and their application to an industrial case study. Using the workflow we manually derive a Communicating Sequential Processes (CSP) model from natural-language safety requirements documents, which Varanus uses as the monitor oracle. This reuse of the model means that the monitor oracle does not have to be developed separately, risking inconsistencies between it and the model for static verification. The approach is demonstrated by the offline RV of a teleoperated manipulation system, called MASCOT, which enables remote operations inside the Joint European Torus (JET) fusion reactor. We describe our model of the MASCOT safety design documents (including how the modelling process revealed an underspecification in the design) and evaluate the Varanus toolchain's utility. The workflow and tool provide validation of the safety documents, traceability of the safety properties from the documentation to the system, and a verified oracle for RV.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes