Understanding the state and action spaces of demonstration data

[Mr-vedant-gupta]

Hi!

I'm trying to understand what exactly the different keys in the image below represent:

Regarding the action space - my understanding is that the first 6 variables represent the desired change in position and orientation of the gripper, while the last variable represents whether it is open or closed. Does this mean that the action space is Markov given the robot state and image of the environment?

Additionally, could you please help me understand what each of the observation keys represents, and what the key "states" with dimension (T, 47 ) represents?

Thanks!