Inconsitent Gazebo sim startup behaviour

[gbg-ivi]

Hello, I am trying to use the Gazebo Simulator which i installed via apt on WSL. I am running into a problem where launching one_drone/tmux/start.sh sometimes works, but most often fails to start the simulator properly. When I restart the script a couple of times it eventually works (so it is possible), but not very consistently. When failing to start properly, tab 3:core repeatedly prints:

[component_container_mt-1] [INFO] [1762786614.487156790] [uav1.estimation_manager]: starting the estimator 'gps_garmin'

[component_container_mt-1] [WARN] [1762786614.487286017] [uav1.estimation_manager]: [gps_garmin/alt_garmin]: Estimator must be in READY_STATE to start it

[component_container_mt-1] [WARN] [1762786615.120830397] [uav1.estimation_manager]: not publishing uav state in INITIALIZED_STATE

[component_container_mt-1] [INFO] [1762786615.123783078] [uav1.estimation_manager]: [gps_garmin/alt_garmin]: WAITING FOR: correction gps_garmin/alt_garmin/pos_garmin

[component_container_mt-1] [WARN] [1762786615.460134459] [uav1.estimation_manager]: range measurement 0.00 outside of its valid range (0.00, 36.00)

[component_container_mt-1] [WARN] [1762786615.634118235] [uav1.estimation_manager]: not publishing uav state in INITIALIZED_STATE

The last message from 2:px4_api is getting magnetometer heading, Tab 5:autostart is stuck on [component_container_mt-1] [WARN] [1762863943.333551562] [uav1.automatic_start]: waiting for data: ControlManager=true, UavManager=true, HW Api=true, EstimationManager=FALSE, tabs 6:takeoff, 7:rviz and 8:layout are stuck on waiting for /uav1/gain_manager/diagnostics

Have you encountered this behavior before or do you know how it could be fixed? Thanks in advance and kind regards :)

[matemat13]

It seems that the garmin rangefinder used for measuring height above ground is giving you zero readings for some reason... is it possible that the drone is spawned inside the ground plane or something? Can you visually check in the Gazebo GUI if the sensor is perhaps blocked by something? Hard to tell what's going wrong from the output you provided and I can't seem to be able to replicate your problem :(

[gbg-ivi]

Thanks for the quick reply :)
I also thought of something like that, but the drone seems to spawn fine. Are there any logs i could provide that might help? The progress map on the core git page has an entry for fixes to the magnetometer for gazebo, is that maybe related to the px4 tab being stuck on getting magnetometer heading?

[spurnvoj]

Hi @gbg-ivi,

The ROS2 version is still a work in progress. We are currently refactoring all sensors and fixing the cameras and lidars to align with the behavior we're accustomed to from ROS1. Perhaps this pull request, #25, will address the issue you're experiencing. Please wait for the merge.

[gbg-ivi]

@spurnvoj running from the sim from branch ft_ros2_rgbdcam of the Gazebo sim repo sadly produces the same behaviour.

I set up WSL on another Windows machine and managed to reproduce the behaviour, maybe its related to that? I'm running Ubuntu 24.04.3

[spurnvoj]

Hi, I have standard Ubuntu 24.04.2, and the system (branch ft_ros2_rgbdcam) works for me. So this could be somehow related to WSL.