Are the heading values of tracks in waymo motion dataset already normalized?

[yanggao0118]

When using the waymo motion dataset, i noticed that some tracks have heading values even greater than 2pi. In the string docs it has mentioned heading should be normalized to [-pi, pi].

Below is the code snippet i used to read the data. I would greatly appreciate it if you could help me identify where the issue might be.

        scenario = scenario_pb2.Scenario()
        scenario.ParseFromString(bytearray(data.numpy()))

        scenario_id = scenario.scenario_id
        ego_id = scenario.sdc_track_index
        
        ego_track = scenario.tracks[ego_id]

        for step in range(len(scenario.timestamps_seconds)):
            ego_state = ego_track.states[step]
            if not ego_state.valid:
                continue

            ego_position = [ego_state.center_x, ego_state.center_y, ego_state.center_z]
            ego_heading = ego_state.heading
            ego_velocity = [ego_state.velocity_x, ego_state.velocity_y, 0]

            for track in scenario.tracks:
                if track.id == ego_track.id:
                    continue 

                object_state = track.states[step]
                heading = object_state.heading

[scott-ettinger]

Thanks for bringing this up. The comment is incorrect in the code. The headings are not necessarily normalized. We will fix the comment.