I built a drone usign the X500v2 kit and a Jetson Orin Nano back here at the lab, where we have a VICON system setup. Ive already made the necesary connections and configuration for the jetson, which is connected though the xrce-dds agent to the Pixhawk 6C, to see and receive the pose estimation at ~118 Hz. After publishing to the /fmu/in/vehicle_visual_odometry topic, with non-specified timestamp and timestamp_sample, I have a pose reassembling the one from the VICON system, but a weird thing is happening, every 10 seconds or so, the topics relating to position and odometry dissappear, and I’ve noticed the EKF2 is not running as fast as without the vision aid, perhaps this could be related to the EKF2_EV_DELAY param? I decided to check on the shell the EKF2 status, and found out that the EKF2 is missing visual odometry events, as seen in the following ouput:
ekf2 status
INFO [ekf2] available instances: 2
INFO [ekf2] 0: ACC: 4259850, GYRO: 4325386, MAG: 396321, healthy, test ratio: 1.0000000 (1.00000)
INFO [ekf2] 1: ACC: 2490378, GYRO: 2490378, MAG: 396321, healthy, test ratio: 1.0000000 (0.00000) *
ekf2:0 EKF dt: 0.0090s, attitude: 1, local position: 1, global position: 0
ekf2: ECL update: 28054 events, 10674us elapsed, 0.38us avg, min 0us max 124us 1.156us rms
ekf2: ECL full update: 28065 events, 2898060us elapsed, 103.26us avg, min 69us max 443us 46.578us rms
ekf2: IMU message missed: 2 events
ekf2: vehicle_air_data messages missed: 0 events
ekf2: vehicle_magnetometer messages missed: 0 events
ekf2: vehicle_visual_odometry messages missed: 5 events
ekf2:1 EKF dt: 0.0099s, attitude: 1, local position: 1, global position: 0
ekf2: ECL update: 25571 events, 10594us elapsed, 0.41us avg, min 0us max 416us 3.157us rms
ekf2: ECL full update: 25571 events, 2723792us elapsed, 106.52us avg, min 69us max 1002us 56.611us rms
ekf2: IMU message missed: 0 events
ekf2: vehicle_air_data messages missed: 0 events
ekf2: vehicle_magnetometer messages missed: 0 events
ekf2: vehicle_visual_odometry messages missed: 6 events
I suspect it might something to do with either packet delay, which Im currently debugging, and the EKF2_EV_Delay param, which ive not tuned yet. Any indications or help would be of great help. (Addittionally, theres a log as an attachment in log_11_2024-5-31-11-05-12.ulg - Google Drive , if that helps).