Open this publication in new window or tab >>2025 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]
Embedded systems deployed in extreme environments must be designed to be robust so that they can operate correctly. For example, satellites must tolerate radiation induced faults, and drones must tolerate strong winds. Designing robust systems is challenging, especially for devices with limited hardware resources and real-time requirements. In this dissertation, we address several problems related to the design of robust real-time embedded systems. We address robustness on two main fronts: ensuring that dynamic software updates are done correctly and at the right time so that any bugs or vulnerabilities can be fixed while the system is deployed, and ensuring that systems can handle faults that occur in extreme environments in a timely manner. We first present a method for scheduling dynamic software updates, which are critical for maintaining and improving the robustness of a system during deployment. Our method schedules dynamic updates while satisfying real-time requirements by scheduling them during times the processor would have been idle. It advances the state of the art by allowing updates to be scheduled without negatively impacting any of the system's real-time tasks. We then present a method of efficiently determining the impact of an update on the timing behavior of the system. Updates that lead to timing violations reduce the robustness of the system and may lead to system failure. We advance the state of the art by exploiting the fact that most updates change very little of the system's code, and hence, we can determine the changes' impact on timing efficiently. We then focus on faults caused by high-radiation environments. To build a robust system in environments where faults can occur, the system must use fault tolerance techniques. However, to know which techniques are suitable, we must build an accurate fault model. We present a measurement platform that quantifies the types of faults and errors that can occur, which form the basis of an accurate fault model. We deploy this measurement platform on a nanosatellite in LEO (Low Earth Orbit). We then present a fault tolerance technique that ensures that errors are handled within a specific time bound, which is critical for real-time systems. Our work advances the state of the art by presenting real-world fault data and by presenting a fault tolerance scheme that can handle long-lasting faults in bounded time, which is vital for real-time systems. The contributions of this dissertation enable designing robust embedded systems.
Place, publisher, year, edition, pages
Uppsala: Acta Universitatis Upsaliensis, 2025. p. 67
Series
Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology, ISSN 1651-6214 ; 2585
Keywords
Robustness, embedded systems, extreme environments, satellites, drones, real-time systems, fault tolerance.
National Category
Embedded Systems
Research subject
Electrical Engineering with Specialisation in Networked Embedded Systems
Identifiers
urn:nbn:se:uu:diva-566110 (URN)978-91-513-2586-6 (ISBN)
Public defence
2025-10-23, Polhemsalen, Ångström, Regementsvägen 10, Uppsala, 09:00 (English)
Opponent
Supervisors
2025-09-292025-09-042026-01-07