Linköping University invites applications for a PhD student position in software engineering for safety and security assurance of autonomous systems fully funded by the CUGS Graduate School in Computer Science at the Department of Computer and Information Science at Linköping University.
Mission: The TestScenS project will provide efficient techniques and tools to improve safety and security assurance for UAV missions. The project aims to automatically generate critical system-level test scenarios for UAV missions from a high-level specification language. The generated scenarios are investigated in existing simulators to validate that no safety or security concerns arise in the presence of various faults or adversarial attacks.
Motivation: Drones are unmanned aerial vehicles (UAVs) with a growing importance in our society. With human authority and various degrees of autonomy, they can easily reach are areas which are unreachable or particularly dangerous for humans (e.g. in search-and-rescue missions or in television broadcast) or unsustainable from a business perspective (e.g. remote monitoring of crops or last mile delivery). However, when drones are operated over populated area, the safe and secure operation of such collaborating agents is of major importance for public safety.
While significant research has been invested to improve the safety of autonomous or human controlled UAVs, concerns for public safety has not yet been justifiably addressed. This project aims to conduct further research to better understand safety and security concerns for UAVs and propose various assurance techniques to provide justifiable evidence for their safe and secure operation.
Scientific Challenges: The TestScenS project aims to investigate the following foundational research questions:
Relevance: The assurance (or certification) necessitates systematic validation independently from the design of UAV agents and missions by investigating that their behavior complies with safety and security requirements inthe presence of different faults and adversarial attacks. As testing is still the most common verification and validation technique used in certification, the project will provide system-level test scenarios as primary artifacts of assurance, potentially combined with formal verification techniques. As a further benefit, our test scenarios can be customized to specific map areas and UAV missions. Finally, the test scenario synthesis tools will be integrated with existing public simulators (such as those available in the WASP Research Arena on Public Safety) to assist the safety and security assurance of UAV missions in adversarial settings.
This project will be supervised by Dániel Varró (main supervisor) and Ulf Kargén (assistant supervisor). Feel free to contact them for further information about the call.