Lighthousedeck LED stay off

[Timo K]

Hello everyone,
I'm pretty new to lighthouse and have some problems with it.
Sometimes the blue LED stays off after blue, red and yellow were on. In another topic someone said this happens, if the firmware image is bad, but I just flashed a new firmware on the Crazyfly using the Crazyradio before I updated the project. So I had the newest firmware and only changed the config file (added CFLAGS += -DDISABLE_LIGHTHOUSE_DRIVER=0) and the positions of the LH Basestations as well as the color of the SYS LED.
I'm using two basestations V1 and changed the distances between them a couple of times and only one Crazyfly 2 at a time with the lighthouse deck.
It worked sometimes with this drone and another one as well, but at the moment they can't find there positions at the beginning, crash when they takeoff or the LEDs on the deck stay off.
Thanks in advance for any help.

[kristoffer]

The correct LED sequence when booting (as of 11/8 2020) is
1. blue + red
2. blue + red + yellow
3. blue

The blue LED should stay on

[kristoffer]

You can check in the console log (in the python client). You should get some log entries that it locked to LH V1 and after a while that it received calibration data

[Timo K]

Thank you for the quick reply. Rightnow we can't reproduce the error with the LEDs so the booting works fine. But there was probably an error when flashing the crazyflie, so the LEDs didnt got to the 3. step.
But we still have problems with the stability of the drone when flying with two v1 basestations. It flies much more stable when only one is turned on and when both are turned on and the crazyflie flies nearer to the second basestation it drops and doesn´t fly in a straight line. This might be caused by a rather big room we are using for the basestations. Currently its about 6x6m big and we are using no sync cable.
Another issue we experienced is that the CF locks to LH2 altough we are using LH1. And that might cause the CF not find its position and therefore doesn´t start. Is it possible to lock it to LH1 (in the firmware)?

[kristoffer]

But we still have problems with the stability of the drone when flying with two v1 basestations. It flies much more stable when only one is turned on and when both are turned on

This sounds like the base station geometry is not OK. If the Crazyflie does not have the correct position and orientation of the base stations, it will have a hard time to calculate its position from the angles it receives from the base stations and will behave badly. With only one base station there is no conflicting data and all is fine.

The base station geometry data is composed of positions and rotation matrixes for the base stations. The rotation matrixes might be tricky to understand, but you should be able to look at the positions to determine if it looks reasonable or not. The positions are relative to the point where the Crazyflie was located when running the get_bs_geometry.py script.

Currently its about 6x6m big and we are using no sync cable.

6x6m sounds large, have you tried a smaller setup and maybe a sync cable? If you get something smaller to work you can try to expand from there to see what causes problems.

Another issue we experienced is that the CF locks to LH2 altough we are using LH1

Yes this is a show stopper. The Crazyflie tries to determine if it is a V1 or V2 system by looking at the pulses it receives, and it usually works pretty well. If it locks to the wrong type of system, the Crazyflie must be restarted, as it will not evaluate the system type again after the lock. If it is locked to the wrong type of system, it will not be able to receive any data from the base stations and positioning will not work at all.
Does this happen consistently or from time to time?

[Timo K]

That happend rather consistently. We had this issue rather often. It was like a 50:50 chance to get the correct LH version. That´s why we changed a function in the firmware of the CF which should detect the LH version to always return LH1 since we currently always use LH1.

But since we only lent the LH1 base stations and would like to benefit of the wider laser angle of LH2 we are planning to buy LH2 base station. In the Lighthouse blog you reported on 10th of april 2020:

We have also written a python script to estimate base station geometry (positions and orientation) using the Lighthouse deck. This removed the requirement to use software from Steam which should simplify the set up process. Please see the (still limited) documentation. Note that this calibration method only supports the basestation V1… for now!

Is the quote still valid or did you also implement the position detection of LH base stations via python for LH2?
It would be a huge advantage for us to get the positions by an python scripts since we move the basestations frequently to other positions.

Thank you for your fast responses.

[kristoffer]

Maybe we need to take a second look at the system identification?

Yes, the script works for LH2 as well.
Note that errors are still larger for LH2 (compared to LH1) since we have not yet implemented corrections for manufacturing (alignment of lenses and so on). This is still work in progress.