Project title: Modeling and Analyzing Cyber-Physical Systems
Research Leader: Marjan Sirjani
Participating Companies over all the Sprints:
- Volvo Cars
- Volvo CE
- Volvo GTO
- DEIF
Vision and Mission
The project explores various aspects of developing Cyber-Physical Systems (CPSs) and autonomous systems. (a) designing new models and languages for the timed and hybrid system. (b) proposing new processes and frameworks for building dependable and resilient CPSs and autonomous systems. (c) implementing new techniques and theories for formal verification and analysis, especially timing analysis. (d) modeling and analyzing the event-based asynchronous autonomous systems for safety and security assurance, performance evaluation, and optimization.
Project domains include collaborating autonomous machines, collaborating agents, intelligent factories, or event-based distributed programs executing across different network nodes.
Goal
The general goal is to structure the overall system requirements, including safety, cybersecurity, schedulability, and performance properties, considering computation time and costs (e.g., energy consumption), and model the behavior of the system based on the architecture and timing features of the system, ensure the safety properties of the system and evaluate the performance and investigate different heuristics for optimization.
The focus of our modeling and analysis is on two features, timing properties and cybersecurity.
To analyze timing, we consider heterogeneous and dynamic systems and develop compositional and scalable methods.
To ensure cybersecurity and cyber-resiliency, we model different attack scenarios, and device the IDS with a monitor to detect cyber-attacks and adapt to the situation.
Achievements
Different tools are developed throughout the project.
End-to-End (E2E) time analysis tool: The tool starts from a component diagram where the system is visualized. The tool presents the End-to-End (E2E) time analysis after getting the period, jitter, offset, and delays of each component.
We are working on a scalable analysis for such systems where we have designs like in a Software Product Lines. We work on covering the TIMMO2 constraints with an interface for EAST ADL models.
AdaptiveFlow http://rebeca-lang.org/allprojects/AdaptiveFlow is a framework that allows to model and analyze track-based flow management systems. In this framework, users provide three input files: environment.xml, topology.xml, and configuration.xml and the tool checks safety and performance.
We published several papers focusing on cyber-physical systems in different venues.
On cybersecurity we published:
“An Actor-based Approach for Security Analysis of Cyber-Physical Systems” at FMICS 2020, and
“Monitoring Cyber-Physical Systems using a Tiny Twin to Prevent Cyber-Attacks” at SPIN 2022.