REFAIRE EN FRENCH --> The SARCoS project will develop intelligent tools for automating the model-based testing of complex connected systems, to find security and reliability problems. By leveraging machine learning techniques, these tools will:
- Infer domain-specific test models
- Generate attack scenarios and by learning from execution interaction data
- Refine and prioritize test cases
- Produce smart analytics of test results
Automated model-based testing will be used to generate security tests targeting system vulnerabilities by false data injection attacks, and robustness tests using behavioral model-based fuzzing.
By automating the most cumbersome parts of the testing process for each target domain with intelligent tools, the project will not kick out validation engineers from their current testing tasks, but rather will help them focus on the design of more complex security and robustness testing scenarios which require manual tuning. The ultimate target of the project is to strongly reduce the startup overheads and costs of testing, to enable cost-effective testing of security-critical and trusted interconnected systems at business level.
The SARCoS project directly involves four industry partners in three distinct application domains:
- Intelligent Parking Systems (IPS) with Flowbird
- Air Traffic Control (ATC) management systems with Smartesting
- Secure Timing (ST) with Gorgy Timing
Combining domain-specific model-based testing and machine learning techniques will allow cognitive automation of security and robustness testing through test model inference, data-driven test selection and prioritization and smart analytics of test results. The investigation and evaluation on three different application domains will strongly help to design common and reusable intelligent technologies for model-based security and robustness testing for detecting business logic flaws of complex connected systems.