[SOLVED]TDOA2 positioning error in high altitude

All discussions related to the Loco Positioning system
Post Reply
john5648
Beginner
Posts: 8
Joined: Sat Nov 28, 2020 7:03 am

[SOLVED]TDOA2 positioning error in high altitude

Post by john5648 »

Hi

i have problem with drone being unstable in high altitude(1.5m and 2m).
In video i send certain coordinate point using send_position_setpoint command to make it hover. (https://youtu.be/3zIm3GidDUo)

I installed LPS using 8 anchors with 8m x 8m space, and anchor height being 0.7m and 2m.

I guessed drone being unstable is because of TDOA2 localization error, so i checked localization std with different altitude.
I tested std of TDOA2 localization in certain point which is the center of LPS, and used tripod to make drone still.
This is the result. Std gets higher with altitude.
std.png
std.png (5.78 KiB) Viewed 9799 times
So i think my guess is right. Now i want to solve the localization error of TDOA2 in high altitude.
I think the problem is UWB antenna of drone facing upward, so if drones gets higher 4 anchors get too low.

Would it work for installing more anchors for example like this? ofcourse using TDOA3.
configuration of node.png
or any good ideas?

Thanks
john
Last edited by john5648 on Sat Jan 23, 2021 3:47 am, edited 2 times in total.
kristoffer
Bitcraze
Posts: 630
Joined: Tue Jun 30, 2015 7:47 am

Re: TDOA2 positioning error in high altitude

Post by kristoffer »

Hi!

As long as the Crazyflie is inside the anchor space (the convex hull) I would not expect the std to change a lot depending on the height. If your top anchors are higher than the Crazyflie, the geometry is identical (from an LPS point of view) to flying close to the floor. It is possible that the floor has some influence but if that is the case I would expect the problems to be close to the floor instead.

It looks like you are using a flow deck as well in the video, that would make a difference though, are your measurements done with or without the flow deck?

Your flying space is 8x8 meters, how high up are the top anchors? If they are at around 2-2.5 meters you get a fairly "flat" geometry which does provide little information (mainly in Z) to the estimator. It is possible that adding an anchor in the top-center of the space would help a bit.

Have you increased the output power of the anchors? That will also make a bit difference in a space of this size.
john5648
Beginner
Posts: 8
Joined: Sat Nov 28, 2020 7:03 am

Re: TDOA2 positioning error in high altitude

Post by john5648 »

Hi kristoffer
Thanks for the reply

Yes i am using flow deck v2. I also checked the ranging std of flow deck and as you can see below flow deck has really good performance.
flow deck std.png
flow deck std.png (4.84 KiB) Viewed 9790 times
This is the picture of how i settled anchors
configuration of anchor.png
below anchors has height of 0.7m~1.19m, and top anchors has height of 2.1m~2.25m.
So flying drone in 1.5m~2m height is in convex hull.

I just increased Tx power and it worked really well wow!
i settled tx power to 30db using set_tx_power.py (https://github.com/bitcraze/lps-node-fi ... x_power.py)

Thanks for the advice!
cfnewbie
Beginner
Posts: 24
Joined: Wed Jun 17, 2020 8:51 am

Re: [SOLVED]TDOA2 positioning error in high altitude

Post by cfnewbie »

May i know how do you generate the STD of xyz positions or log them? thank you!
Post Reply