Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Is there a guideline to easily and reliably configure the various weights of the evaluation function? #89

Open
JustMrZuo opened this issue Sep 27, 2023 · 0 comments

Comments

@JustMrZuo
Copy link

Hi there. I find it a very interesting idea to achieve good tracking of a predetermined route based on vehicle dynamic and model predictive control algorithm, which eventually transforms the motion planning problem into mathematical quadratic programming. Therefore, I am writing a simplified version based on your open-source code to achieve my own control goals.

But along the way I encountered the question of whether there exists a guideline to easily and reliably configure the various weights of the evaluation function in the MPC algorithm.

The existing code directly preconfigured the weights of the evaluation function, but this also leads to two problems. On the one hand, it is difficult to visualize the impact of weight changes on the actual control effects, or how sensitive the actual control effects are to particular weight changes. On the other hand, after changing the dynamical parameters or equations of the control object, it is obvious that the formulation of the weights needs to be restarted, which involves the choice of how to take the initial values.

Currently, I'm trying to log and visualize the items in the STAGE parameter in a physical sense, with the expectation of being able to perform a more intuitive analysis based on this. However, I don't have much optimistic expectations for this solution idea at the moment.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant