Original title: Real-Time Model Predictive Control Algorithms for Autonomous Spacecraft Guidance
Mohammed-Adnane Garab · September 2026
Guiding a spacecraft to safely meet another in orbit is difficult because orbital dynamics are tricky and computers must make decisions in real time. Researchers compared four types of Model Predictive Control, which are algorithms that repeatedly predict future motion to steer a vehicle. Using an orbital relative-motion model, they discovered that long prediction horizons make math problems much harder for gradient-based solvers like Fast MPC, causing them to fail, while exact solvers using Cholesky factorisation succeed. Tube MPC, a version designed to handle uncertainty, successfully guided five different rendezvous manoeuvres with one fixed setup. It also kept tracking errors safely bounded, even when following complex, arbitrary flight paths updated in real time. Because its custom-built code matched standard scientific tools to machine precision, Tube MPC offers a realistic, robust choice for onboard flight computers.
Packing a suitcase for a flight, you aim to use all the space without exceeding the weight limit. But if you add items without checking, you might only realize you're over when it's too late. Here, the system plans movements ahead while keeping the spacecraft within safe error bounds, so it never drifts off course during maneuvers.
This helps spacecraft navigate and dock without constant commands from Earth, making missions safer and more autonomous. Engineers can trust the system to handle complex maneuvers in real-time.
The abstract doesn't explicitly state limitations, but the study is confined to the Clohessy-Wiltshire-Hill model and standard rendezvous maneuvers, so it may not generalize to all spacecraft dynamics or environments.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Real-Time Model Predictive Control Algorithms for Autonomous Spacecraft Guidance — Mohammed-Adnane Garab
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.