Help with Stability Issues in Lighthouse Positioning System

[dakeprithvi]

Hi!,

We recently transitioned to the new Lighthouse positioning system. While most forum discussions suggest the basestation can capture a volume of up to 5x5x2 meters, we are working in a smaller space (3x2x2 meters).

Here’s our setup:

• Basestations:
  Mounted on tripods at two corners of the side parallel to the y-axis, angled downwards at 45 degrees.
  Within the 3x2x2 volume.
• Observations:
  This arrangement provided the most stable results compared to other setups (e.g., diagonal placements).
  However, the area of stable operation appears much smaller than 3x2x2.
  If the drone moves ~0.5 meters away from the origin, it sways significantly and crashes.
  Within the 3x2x2, there are also signal drops in certain spots during geometry estimation.

Here's a video highlighting the issue: Video link. Timestamps 0:25 and 0:35 seconds (sway before we ask it to go back to origin).

Despite trying multiple geometry setups, the current arrangement gives the best results, but it’s still far from what we need.

Could anyone help explain:

• Why signal loss occurs even within the 3x2x2 space? (We took care to put window blinds to prevent sunlight)
• Any advice on resolving these issues or improving the setup? (Tried multiple setups, with more height > 3 m, still no results)

Thanks in advance for any insights or suggestions!

[ArisMorgens]

Hello! Did you follow the getting started guide?
Make sure that both base stations are securely mounted in a very stable position as even slight movements of a base station after the geometry estimation can lead to system instabilities.
Also, try to avoid reflective surfaces in your flight area, like the chairs shown in the video you sent.
Finally, the red light on your Crazyflie indicates a low battery level, which can also cause instability in its performance.

[dakeprithvi]

Ok we moved our setup to a more empty, dark room. Also the tripods are placed at 2.5 meters height, and we are operating 50 cm below them.

Before starting geometry calibration, we conducted some experiments and noticed the following issues:

• The drone often fails to receive signals from one or both basestations, or the signals are intermittent, as observed in the Lighthouse tab of the cfclient. Even slight changes in the drone's orientation can cause it to lose signals, even in positions where it previously received them.
• For example:
  say, if oriented along the x-axis, the drone consistently receives signals from both basestations.
  while along the y-axis, it frequently loses signals from both.

This is extremely frustrating because it seems that, no matter how well we perform geometry estimation, the signals are already faulty from the start.

Has anyone experienced similar issues? What could be causing this, and how can we fix it?

Thanks in advance for your help!

[ArisMorgens]

Hello! I just checked again your video and it seems like you are using the long pin headers (male long deck connectors) to attach your lighthouse deck on the Crazyflie. You can try to use the short ones so they don't block the laser sweeps from the base stations.

[dakeprithvi]

Thanks for the suggestion. Changing the connectors did some magic!

We could also perform the autonomous sequence experiment we wanted, and the results are quite better than the previous Loco-positioning system.
Here's a video if interested:
Crazyflie_with_MPC_trajectory_tracking

Thanks again,
Prithvi