search

osrf / vrx

Virtual RobotX (VRX) resources.
Apache License 2.0
433 stars 194 forks source link

IMU and GPS data corrupted? #789

Closed spsingh37 closed 1 week ago

spsingh37 commented 3 weeks ago

I'm testing IMU-based localization (pure integration) and GPS-based trajectory in ROS2 for the WAM-V robot, but I'm noticing that the two generated trajectories do not match the ground-truth odometry, as if there's some scaling issue?

I've set both the bias and noise parameters to zero in the wamv_imu.xacro file to simulate an ideal IMU. After this, I ran colcon build --merge-install to ensure the changes took effect. But the GPS is still pristine (didn't modify its noise).

This has left me questioning whether the IMU data and GPS data themselves might be corrupted or if there’s another issue at play. Please help me why the trajectories differ

To provide more context, I'm attaching a few plots comparing the trajectories with the ground-truth odometry.

Figure_1

spsingh37 commented 1 week ago

I'm closing this issue because the reason for the IMU derived trajectory being underscaled is inconsistent time between consecutive IMU messages. And the GPS is overscaled because the conversion from GPS coordinates to x-y positions was incorrect (I forgot to account for the Earth's shape and instead did a 2D projection)