rockets

« Back: Results

References

To complete this project, I relied primarily on the following sources of information, as well as others I linked to throughout the website:

Textbooks

Online

Graduate Theses and Academic Papers

In particular, the paper by Ferrante was very helpful. We essentially solved the same problem, but with different techniques.

Future Work

I have a lot of work planned for this project. Some of the highlights include:

  • Implementing a more robust method of selecting cost function weights for the LQR controller and using this to compute the flight of a Falcon 9 rocket
  • Implementing model-predictive control (MPC) to replace the LQR controller; this will reduce some of the approximations made in the LQR controller and allow for a more optimal solution
  • Incorporate noise, errors, perturbations in the simulation, such as wind/aerodynamics, state measurements, c.g. offset, etc. and determine additional methods of making controller more robust
  • Studying the use of continuous-time reinforcement learning algorithms in solving this problem

Copyright © 2024 ryan@freestatelabs