Closed kavikode closed 2 years ago
In general, you can test out the arm manipulation environment using the command
python3 -m assistive_gym --env "ArmManipulationSawyer-v1"
The figure you linked to is an example in which human preferences are disabled. This is done by commenting out this line: https://github.com/Healthcare-Robotics/assistive-gym/blob/main/assistive_gym/envs/arm_manipulation.py#L28 and then training a new robot control policy (see the Wiki: https://github.com/Healthcare-Robotics/assistive-gym/wiki/5.-Training-New-Policies).
The human arms have two types of joint limits. The first is pose-independent hard limits, which are set here: https://github.com/Healthcare-Robotics/assistive-gym/blob/main/assistive_gym/envs/human_creation.py#L217 The second type of joint limits that are applied to the arms are pose-dependent limits (put your hand behind your back and you will likely notice it is more difficult to fully bend your elbow). Pose-dependent joint limits are applied here: https://github.com/Healthcare-Robotics/assistive-gym/blob/main/assistive_gym/envs/agents/human.py#L129
For Google Colab, you can edit this notebook to use the environment ArmManipulationSawyer-v1
https://colab.research.google.com/drive/1PAY5HnLKRB-TBsPaevRr6myMfpVt_yzF?usp=sharing
Here is an example of training a new policy in Google Colab. https://colab.research.google.com/drive/1P6zYEL2Pfq7GOCtmoLlRNGrKkPQTVAQF?usp=sharing
Thank you for your messages. How can we use this Assistive Gym Arm Manipulation example in ROS?
Can you please help me understand how to run the arm manipulation task shown in the right side, link: https://github.com/Healthcare-Robotics/assistive-gym/blob/9a308a8a989f179f1259824b579a063d610df110/images/human_preferences.gif?
How can we change the joint limitations of the human model for arm manipulation?
Thank you!