Using SLAM in evaluation?

[DannyLi001]

Hi, thank you for your contribution. I have a question about the evaluation part. How do you get the current qpos? Are you keep doing SLAM or is it read from the robot API? If it's the first case, how do you handle the time offset?

Thank you