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

CPU Board attached to the Bebop drone #19

Open
mandelyoni opened this issue Oct 10, 2018 · 15 comments
Open

CPU Board attached to the Bebop drone #19

mandelyoni opened this issue Oct 10, 2018 · 15 comments

Comments

@mandelyoni
Copy link

Hi,

Which CPU Board is recommended to attach to the Bebop drone for running the net?
If available, please share a link.

Thank you.

@antonilo
Copy link

Hi,
We are using the Intel UpBoard (https://up-shop.org/27-up) for onboard computation.

@mandelyoni
Copy link
Author

Thank you for the prompt response.
By the way, are you going to sell the small docking board with Dronet for the crazyflie soon?

@antonilo
Copy link

We do not plan to sell it since we are still in development phase. While the board can be already purchased, the integrated system will follow in the (hopefully) near future.

@mandelyoni
Copy link
Author

Hi, thanks for your answer.
Which board, did you mean, can already be purchased?
Is it straight forward to implement the the Dronet on it?

@antonilo
Copy link

The PULP board that runs dronet in Ultra-low power can already be purchased. Implementation of the algorithm that is highly non-trivial, but we are currently on it. We will soon share our findings and implementation.

@mandelyoni
Copy link
Author

Hi Antonilo,

Can you please share an explanation/description of how to connect the Up-board to the Bebop2?

Thanks, Yoni

@antonilo
Copy link

I would argue that the easier way to do it is using ROS. This is however not the only one and for sure not the best one in term of performance.

@mandelyoni
Copy link
Author

Thanks, but I meant how connect the Up-Board physically to the drone (power and communication)?

Thanks again, Yoni

@antonilo
Copy link

We have an external battery to power the Up-Board and communication happens via WiFi and ROS.
Hope this helps!

@mandelyoni
Copy link
Author

Hi Antonilo,

Is there a big difference in the performance between running the algorithms on the UpBoard or on a laptop ?
In most of your videos we didn't see the UpBoard connected to your drone? Can you please explain why?

Thank you, Yoni.

@antonilo
Copy link

In the original dronet video, computation was done off-board in a laptop. In our last projects, all computation has been shifted to fully on-board: Deep Drone Racing, Beauty and the Beast. The minimum inference time required to give good behavior is generally 8-10Hz.

@mandelyoni
Copy link
Author

Hi Antonio,
In your paper, DroNet: Learning to Fly by Driving, in Table 3, you make a comparison with Gandhi et al.
Did you also implement their algorithm?
Thank you, Yoni

@antonilo
Copy link

antonilo commented Dec 2, 2018

Yes, we've re-implemented their algorithm as described in the paper.

@mandelyoni
Copy link
Author

Ok, thank you.

@mandelyoni
Copy link
Author

Hi Antonio,
Did you run Ubuntu 16.04 with ROS on the Upboard?
Is the simple version of the board adequate (https://up-shop.org/up-boards/272-up-board.html)?
Thanks, Yoni

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

2 participants