search

roboterax / humanoid-gym

Humanoid-Gym: Reinforcement Learning for Humanoid Robot with Zero-Shot Sim2Real Transfer https://arxiv.org/abs/2404.05695
https://sites.google.com/view/humanoid-gym/
705 stars 117 forks source link

Domain rand #2

Closed ChengEeee closed 6 months ago

ChengEeee commented 6 months ago

I see that the code has many domain randomization Settings, as shown in the figure, but only a small part of them are used. Can such Settings ensure the robustness of sim2real image image

zlw21gxy commented 6 months ago

We have identified and resolved some unused settings in the file. Additionally, we are excited to introduce a new DR term called 'Dynamic Randomization.' Before transferring your policy to the real world, we recommend using the 'sim2sim' tool. We have also enhanced the sim2im feature. You are welcome to try these enhancements and provide your feedback.