Bitcraze Forums

I am having trouble using two basestations (v2). Depending on which basestation is used I get vastly different results.
When I am plotting lighthouse.angle0x the values change between +-0.025, and when plotting lighthouse.angle0y, the values change between -0.25 and -0.65.
If I cover/block one of the basestations the plots are smooth and reasonable but as soon as I let the other basestation take over the readings shift in an instant.

I followed all the steps of the setup process with calibration data and geometry data , but somewhere it must have gone wrong.

Hi!

Have you configured the base stations to use channels (modes) 1 and 2? This part was unfortunately missing in the instructions but was added recently. You can find the instructions here https://github.com/bitcraze/crazyflie-f ... hthouse-v2

The other thing to check is that your calibration data is correct. You'll find instructions in the same page of how to extract it from the base stations.

Thank you. I had not set the channels and that seem to have solved the problem of the switching values.

Not sure if this falls under the same issue, but I will post it here anyways.
I now get lighthouse.angle0x and lighthouse.angle0y values that reads smoothly.
However, the values does not fully correspond to the XYZ-frame based on the CFs origin, my thinking about this is that the read lighthouse.angle0x values correspond to the basestation reference frame.

Next, as soon as I manually lift the CF into the air (without covering the Lighthouse deck) the state estimates starts jumping around. This I have no idea about and would appreciate some help.

It sounds like your geometry data may be wrong, have you set it according to the instructions?

We are currently adding lighthouse configuration functionality to the client, and even-though we're not done yet, we have implemented a visualization of the system. If you update the firmware, python lib and the python client to the latest versions from github, you will find that there is a lighthouse tab with a 3D system view. It might help you to understand if the geometry data looks OK.

Now it works! I thought that I had done the geometry part, but apparently not correctly.
Perhaps it could also have been that previously when I got the geometry data I had placed the CF in the wrong direction?

Anyways, thank you for the help!