 # Transforming the ego_pose coordinates into global coordinate frame

Hi,
Is there a way to transform the coordinates from the ego_pose information into the global coordinate frame? Thank you very much for the answer.

Yes, that’s what the ego_pose table is there for.

``````    # Get reference pose and timestamp.
sample_rec = nusc.sample
ref_chan = 'LIDAR_TOP'
ref_sd_token = sample_rec['data'][ref_chan]
ref_sd_rec = nusc.get('sample_data', ref_sd_token)
ref_pose_rec = nusc.get('ego_pose', ref_sd_rec['ego_pose_token'])

# Homogeneous transformation matrix from global to _current_ ego car frame.
car_from_global = transform_matrix(ref_pose_rec['translation'], Quaternion(ref_pose_rec['rotation']),
inverse=True)``````

Hi,

the ego_pose information is already given into the global coordinate frame, isn’t it?

See tutorial 11. ego_pose

``````ego_pose contains information about the location (encoded in translation) and the orientation (encoded in rotation) of the ego vehicle body frame, with respect to the global coordinate system.
``````

Tank you for your answer Hi. That sentence may sound a bit confusing.
What it really means is

ego_pose contains information about the location (encoded in translation) and the orientation (encoded in rotation) of the ego vehicle, with respect to the global coordinate system.

Hi,

So I want to know, how fast the ego-car ist moving at timestamp t1: Then I need the yaw-rate (angle °) of the ego-car, I am calculating it like: Is it right or do you have another way to calculate the ego_velocities from the ego_pose?
I am asking because I want to compare the ego_velocity to the radar-measurements

Best regards

Why not simply use your ego_velocity formula?
Also note that the radar provides both compensated and uncompensated velocities (w.r.t. ego motion).