Skip to content
BranchesTags
 
 

Repository files navigation

Tutorial for Isaac Lab Projects

This repository is a tutorial repository, designed to go hand in hand with the Getting Started walkthrough found in the Isaac Lab docs. Each branch of the repository represents another stage of the walkthrough, with the main branch being the "final state" of the tutorial project.
$$% 1. Observation (Naive) \mathbf{obs} = \begin{bmatrix} \mathbf{v} \\\ \mathbf{c} \end{bmatrix}, \quad \dim(\mathbf{obs}) = 6 + 3 = 9 % 2. Forward vector \mathbf{f}_{world} = q_{root} \otimes [1, 0, 0] % 3. Reward (Naive) r_{total} = r_{forward} + r_{alignment} \quad r_{forward} = v_{b,x}, \quad r_{alignment} = \mathbf{f}_{world} \cdot \mathbf{c} % 4. Observation Tuning \mathrm{dot} = \mathbf{f}_{world} \cdot \mathbf{c} \quad \mathrm{cross}_z = (\mathbf{f}_{world} \times \mathbf{c})_z \quad v_{forward} = v_{b,x} \quad \mathbf{obs} = \begin{bmatrix} \mathrm{dot} \\\ \mathrm{cross}_z \\\ v_{forward} \end{bmatrix} % 5. Reward (Tuned Multiplicative) r_{total} = r_{forward} \cdot r_{alignment} \quad r_{forward} = v_{b,x}, \quad r_{alignment} = \mathbf{f}_{world} \cdot \mathbf{c} % 6. Reward with Exponential Alignment r_{total} = r_{forward} \cdot e^{r_{alignment}} \quad r_{alignment} = \mathbf{f}_{world} \cdot \mathbf{c}, \quad r_{forward} = v_{b,x}$$

About

Tutorial project for the accompanying doc

Resources

Readme
License

Contributing

Contributing
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages