Skip to content

Conversation

@suren-atoyan
Copy link

Thanks for the great article 🙌

Copy link
Author

@suren-atoyan suren-atoyan left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Screenshot 2025-02-09 at 01 31 40

This section is missing from the README file 🤔

Screenshot 2025-02-09 at 01 32 28

This section is also missing from the README file and there is an extra period at the end.

The robot is encouraged to maintain a desired height as specified by the commanded altitude. A penalty is applied for deviations from this target height:

$$
R_{z} = (z - z_{ref})^2
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This was a bit confusing, as the reward should be bigger when the robot is at the desired height, and smaller when it deviates from it. So, I think, we either need to change the formula (e.g. $\exp[-(z - z_{ref})^2])$ or keep the same formula but change the name of the variable to "penalty" instead of "reward" (basically, what the title suggests). I go with the second option, but I realize that it might add some confusion around the naming of the variables. Let me know what you think.

- $z_{ref}$ is the target height specified in the commands.

#### 4. **Pose Similarity Reward**
#### 4. **Pose Deviation Penalty**
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here, I also changed the title to keep the consistency with the formula.

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

Successfully merging this pull request may close these issues.

1 participant