Crazyflie does not fly properly

Post here to get support
whoenig
Expert
Posts: 355
Joined: Mon Oct 27, 2014 2:55 am

Re: Crazyflie does not fly properly

Post by whoenig »

The ROS topic for /cf2/log1 uses a generic topic and is not aware of which variable you are trying to inspect. The proper way of visualizing is:

Code: Select all

rqt_plot /cf2/log1/values[0]
The order in the array matches the order of your configuration.
santiagorg2401
Member
Posts: 41
Joined: Sun Sep 08, 2019 2:28 pm

Re: Crazyflie does not fly properly

Post by santiagorg2401 »

Hi, thank you, I can only go to the lab on monday, so I will try next week and post the graphs, another problem I saw is that when I placed the Cf near the origin (LPS Node 0) (say a distance of 20 cm or less), it totally lost the position estimator, for example, I place it 20 cm away of origin (inside the hull) and the state estimator shows a position of -30 cm (out of the hull).
santiagorg2401
Member
Posts: 41
Joined: Sun Sep 08, 2019 2:28 pm

Re: Crazyflie does not fly properly

Post by santiagorg2401 »

Hi,
I am sorry I did'nt upload the plot, I was resetting the flight space bigger as you recommended, now It is 2.89 by 2.89 by 2.0 meters [x,y,z], I couldn't add more length to Z because it was already 20cm away to the roof, here it is the plot of the state estimates x,y, and z in the video in this link https://drive.google.com/file/d/1DLzMMR ... sp=sharing.
santiagorg2401
Member
Posts: 41
Joined: Sun Sep 08, 2019 2:28 pm

Re: Crazyflie does not fly properly

Post by santiagorg2401 »

It gets really noisy near to anchor 0, as you may see in this video https://drive.google.com/file/d/1Mb97m0 ... sp=sharing, the CF is still in the coordinates [0.40, 0.30, 0.00]
The stability improves a bit when the CF is in [0.50 ,0.50, 0.00], but overall when the cf is in the near of anchor 0 and is on the floor, it is not very stable.
I have a doubt, the anchors's Z positions that are near the floor (15 cm away) are supposed to be 0.0 or 0.15? (in the anchor configuration in cfclient)
kristoffer
Bitcraze
Posts: 390
Joined: Tue Jun 30, 2015 7:47 am

Re: Crazyflie does not fly properly

Post by kristoffer »

I recently discovered that the antenna delay seems to be wrong (https://github.com/bitcraze/crazyflie-f ... issues/680) and this causes the distance measurements in TWR to have a negative offset of around 20-30 cm. Maybe this is what you are seeing?
santiagorg2401
Member
Posts: 41
Joined: Sun Sep 08, 2019 2:28 pm

Re: Crazyflie does not fly properly

Post by santiagorg2401 »

Hi,
I am using TDoA v2, I do not know if that same problem applies with it too.
kristoffer
Bitcraze
Posts: 390
Joined: Tue Jun 30, 2015 7:47 am

Re: Crazyflie does not fly properly

Post by kristoffer »

The antenna delay is cancelled out in TDoA and should not have any effect.
I don't know why you get noisy results, but you mention that it is worse close to the floor, maybe the floor is causing problems? If there is a lot of metal in the floor (rebar, sheet metal or similar) it might act as a mirror.
santiagorg2401
Member
Posts: 41
Joined: Sun Sep 08, 2019 2:28 pm

Re: Crazyflie does not fly properly

Post by santiagorg2401 »

No, the floor is ceramic and it has a plastic mat on top, this is a photo of the lab I took several days ago when I was extending the box [please check attachments], of course when working those boxes are not there, and it is at least 30 cm away from any metal structure, the white tubes you see are made of PVC, I do not have a newer picture but currently the mat is more organized, the cables are cut at a shorter length, and the box is centered so there is enough space from the sides.
Attachments
WhatsApp Image 2021-02-19 at 16.09.18.jpeg
Post Reply