Quadruped-PyMPC
Quadruped-PyMPC copied to clipboard
iit-DLSLab
A model predictive controller for quadruped robots based on the single rigid body model and written in python. Gradient-based (acados) or Sampling-based (jax).
Hey guys as mentioned yesterday this is the pull request I'm reopening for the controller class. The only thing missing is fixed the sampling I will try to get it...
We need to use the control instances outside of the repository in a modular fashion. - [ ] Instantiate the controllers without using the config file after the initialization. I.e.,...
For ease of use, the robots should be loaded from robot_descriptions.py. If custom modifications to the XML are needed, I believe all can be done programmatically instead of handcrafting the...
Users should be able to introduce a new robot, by defining a .yaml config file defining the robot name, and the necessary attributes to be handled by the env.
Most modules in this repository import a fixed repo-specific config file, config.py, and use it to load class-specific and function-specific attributes. This practice poses a problem, as external users must...
This commit removes the multiple imports to the configuration file from different modules within the pipeline. This hinders the hability to customize the system's configuration, since each submodule is reloading...
# Dataset generation pipeline using the simulation script. This commit introduces changes to the simulation script so it can be used for generating the datasets of sensor observations from controlled...
Hi, Thanks for open-sourcing the code. I'm trying to run a quadruped in mujoco with the Visual Foothold adaptation. I'm trying to get it working for stepping stones like scenarios...
The check_touch_down_condition() method in SwingTrajectoryController had incorrect logic that detected lift-off events instead of touchdown events. This caused step frequency optimization to fail with slow gaits, particularly crawl patterns at...
Metadata
Owner
Metadata
A model predictive controller for quadruped robots based on the single rigid body model and written in python. Gradient-based (acados) or Sampling-based (jax).