Render the ground truth trajectory for CAM_FRONT

[arvindks89]

Given an instance, I see this method to predict the future trajectory. But what coordinate system is the output ?
And how do I render it on CAM_FRONT image for example.
get_future_for_agent(instance_token, sample_token, seconds=6, in_agent_frame=True)
Ref: UniAD/projects/mmdet3d_plugin/datasets/data_utils/trajectory_api.py
Any hint will be very helpful