Real-Time Optimization Strategies for Predictive Motion Control in Space Robotics

Master’s Thesis / Internship Real-Time Optimization Strategies for Predictive Motion Control in Space Robotics (f/m/x)

Space robotics is an emerging field of growing relevance for applications such as on-orbit servicing, on-orbit assembly, and active debris removal. In these domains, robotic motion and interaction take place in highly constrained environments and under strict computational and safety requirements. Recent research has shown that Model Predictive Control (MPC) can provide perfromant motion control while explicitly accounting for constraints and uncertainties.

Robotic control problems are inherently nonlinear. Therefore, applying Nonlinear Model Predictive Control (NMPC) in real time requires an efficient formulation of the optimal control problem, together with a carefully selected discretization and solution strategy. Previous work has investigated an offline optimization procedure for identifying the most effective discretization methods. This thesis will investigate convexification methods to improve runtime performance on space-representative hardware while maintaining satisfactory closed-loop performance.

Your Contribution

We are seeking a motivated master’s student with an interest in space robotics to implement an onboard model predictive controller.

Your Tasks Implement an existing NMPC controller on space-representative real-time hardware. Investigate successive convexification strategies for nonlinear operational constraints. Derive bounds on the model mismatch introduced by convexification. Benchmark runtime performance and closed-loop behavior on real-time hardware. Your Qualifications Currently enrolled in a master’s program in aerospace engineering, mechatronics, computer science, robotics, mathematics, or a related field. Strong programming skills in C/C++; experience with real-time or embedded implementation is a plus. Working knowledge of MATLAB or Python is preferred. Strong interest in space robotics or spacecraft dynamics. Familiarity with Model Predictive Control and numerical optimization is an advantage. We Offer The opportunity to work on a real-world research problem relevant to autonomous space missions. Collaboration within a leading research institute in space robotics. Insight into robust model predictive control for safety-critical systems.

This thesis is ideal for students interested in embedded optimization, numerical simulation, and space applications who enjoy translating advanced algorithms into efficient, high-performance implementations.

Supervisors: Peter Kötting and Roberto Lampariello