Hello,
Task: Robotic Manipulation using RGB images and joint state information
I am trying to setup a custom RL Algorithm (PPO) using the Manager based environment in IsaacLab. I am using skrl library for its PPO architecture, and the inputs are camera images and robot joint states. I wanted to understand if any examples within IsaacLab utilize camera and state information both as inputs.
I read the skrl_camera_ppo_cfg.yaml for the cartpole example with camera images as inputs, but I did not understand how I could combine state information in that as well.
Any kind of advice or resources would be helpful. Thank you