Parameter Synthesis and Requirement Analysis for Cyber-Physical Control System Design

Projected started in 2018 and is ongoing.

Current design processes for cyber-physical systems mostly rely on extensive testing to ensure the final product complies with its specifications. However, design errors discovered towards the end of design cycles are costly both financially and in terms of time. A model-based approach that uses mathematical models of the system dynamics and formal specifications capturing requirements and assumptions, and that is supported by formal verification and control synthesis algorithms can be extremely useful to discover inconsistencies early in the design cycle and to provide guarantees on the closed-loop behavior. Such requirement driven design processes are particularly useful for system integrators like GVSC as they coordinate different design teams and integrate subsystems from different subcontractors.

This project aims to develop methodologies to understand how requirement satisfaction and safe operating envelopes of systems depend on certain parameters of the system design or external environment, including the parameters of other subsystems the system interacts with. The proposed quantitative methodologies and tools will enable GVSC engineers to analyze requirements and study the effects of different parameters in requirement satisfaction.

The overarching objective is to provide requirement-driven principled design methodologies for cyberphysical control systems. Our focus is on functional requirements related to the dynamical behavior of the systems. We develop algorithms to automatically synthesize controllers for systems together with safety envelopes for these controllers that allow us to do design-space exploration by changing parameter ranges. The developed algorithms and tools will help identify the trade-offs between parameters and individual requirements.

Publications and Presentations

• Invited Talk, “Towards correct-by-construction controller synthesis for self-driving cars”, 7th Workshop on Synthesis (SYNT 2018), Oxford, UK, July 2018.