Future space missions become increasingly challenging following the introduction of re-usable launchers and concepts of manned mars landing. Due to the increasing system complexity and autonomy, classical methods for planning, guidance, and control are no longer adequate to assure mission safety and success. Re-usable launcher and landing systems have rapidly varying dynamics due to the high fuel fraction, rapid fuel-burn, and changing environmental conditions. This must be explicitly considered by the control and planning strategies. The state-of-the-art approach is heading towards onboard, online optimization to meet the performance and autonomy requirements. However, standard real-time embedded optimization methods need convergence guarantees when dealing with nonlinear dynamics under uncertainties. While formal Validation and Verification (V&V) tools are well elaborated for linear systems, they are missing for nonlinear online-optimized systems. Hence, the project aims at developing analytical V&V tools to explicitly meet the requirements for uncertain nonlinear systems under safety constraints over finite time and control horizons. The existence and fixed-point convergence will be guaranteed using the integral quadratic constraints (IQC) framework. IQCs allow to systematically assess the impact of uncertainties and nonlinearities in nonlinear systems via efficient optimization. Using a probabilistic approach to IQCs significantly reduces the analysis's computational effort. Ultimately, this allows for the online robust performance and stability monitoring of onboard optimization, planning, and control strategies. Furthermore, the activity shall develop online safety filters using control barrier functions preventing unintended excursions of the system. Finally, the proposed robustness monitor's applicability shall be demonstrated via flight test on the ADAMP testbed. Hence, the work contributes completely novel analytical online V&V tools for space systems.
