Model predictive control (MPC) is a discrete-time multi-variable control architecture. It uses an internal model to predict the future behavior of the system that is being controlled, in this thesis, a spaceship. Based on the predictions, the optimal control actions are taken. MPC is designed to stabilize a system along the desired path while fulfilling its physical constraints. In automated driving applications, MPC is used as adaptive cruise control, assisting lane-keeping, lane-following control, parking and obstacle avoidance. In general, controlling the vehicle improves responsiveness while maintaining passenger comfort and safety.

Kerbal space program is a space flight simulation video game in which players can create rockets, aircraft, and other craft from a provided set of components. The players control their spacecraft in three dimensions and the game has been praised for its accurate orbital mechanics. All objects in the game except the celestial bodies are simulated using Newtonian dynamics thereby depicting realistic scenarios. A spacecraft has the ability to enter orbit or even travel to other celestial bodies. Certain properties such as the orbit or trajectory of the player vehicle, as well as the position and trajectory of other spacecraft and planetary bodies can be displayed to give an idea of the trajectories.

This thesis describes the use of model predictive control for the landing of a spaceship simulated in the "Kerbal space program". After introducing the theoretical concepts and the simulation setup, the model for the spaceship is constructed and explained. Finally, using these concepts, the control policy for landing and stabilizing spaceship is derived.