google-deepmind / mujoco_mpc

Real-time behaviour synthesis with MuJoCo, using Predictive Control
https://github.com/deepmind/mujoco_mpc
Apache License 2.0
982 stars 148 forks source link

Task::ModifyState #261

Closed thowell closed 8 months ago

thowell commented 8 months ago

Adds functionality to Task to modify State object before use by planner. Use case: experimenting with noisy state estimates.

To demonstrate a use case, the time-varying particle tracking task is modified to enable Gaussian noise injection to the state.

258 @alberthli