jc-bao / policy-adaptation-survey

This repository is for comparing the prevailing adaptive control method in both control and learning communities.
Apache License 2.0
7 stars 1 forks source link

Hovering finally converge to a point close to origin given no external uncertainty. #28

Closed jc-bao closed 1 year ago

jc-bao commented 1 year ago

image

jc-bao commented 1 year ago

Assumption The penalty for velocity is too large

Trail try different reward function

Result Now the control error becomes smaller.

🧑‍🏫 Lesson Tracking and Hovering are different tasks, and should use different reward functions.