Issues in example_ekf_elfo_gps.cc Preventing EKF Convergence

[jijushifei1]

Dear author,

I hope this message finds you well.

I am currently working with the LuPNT example example_ekf_elfo_gps.cc, but I am encountering several issues that appear to prevent the EKF from converging. After reviewing the code and running the example, I believe there are multiple problems in the setup and timing logic that may be affecting the simulation consistency and the measurement generation.

In particular, I observed the following potential issues:

Epoch initialization / TLE loading order

The code queries the constellation epoch (GetEpoch()) before the TLE file is loaded. This may result in an uninitialized or incorrect epoch0, which then propagates through the entire simulation timeline.

Constellation/channel/device binding order

The GNSS constellation may not be properly connected to the channel and devices if the channel is not created and assigned before loading the TLEs. This can lead to missing transmitters in the channel, and ultimately no valid GNSS signals reaching the receiver.

In my runs, the tracked satellite count was often zero or inconsistent across steps, which would naturally prevent the EKF from receiving enough information to converge.

Because of these issues, I am unable to obtain stable measurements and the EKF does not converge as expected. I would be very grateful for your guidance on the intended time/epoch handling and the proper initialization sequence (channel → constellation → receiver/transmitter → epoch → propagation/measurement). If you have a recommended patch, reference implementation, or updated example, I would sincerely appreciate it.

Thank you very much for your time and assistance. I respect the effort that went into developing this library and would be grateful for any help you can provide.

With kind regards,
jijushifei1