MATLAB version of safe control with supervised learning
requirement: OSQP (MATLAB version)
Running instruction
% go to the project path RetrofitAI-matlab
install_osqp % install the requirement
addpath("translated matlab")
run test_adaptive_clbf.mThe original repo: https://github.com/ren365/RetrofitAI
Reference paper: https://arxiv.org/abs/1910.02325
The structure of the folder.
| Name | Type | Description | Status |
|---|---|---|---|
| original python | folder | [test] original python files | done |
| osqp | folder | installed lib for QP solver | done |
| test codes | folder | [test] test whether matlab & python codes behave the same | done |
| translated matlab | folder | [main] matlab files | done |
| install_osqp | .m | automatically install the osqp on matlab | done |
Inside the "translated matlab" folder, the descriptions of files are shown in the table below.
| Name | Type | Description |
|---|---|---|
| test_adaptive_clbf | .m | main file to run |
| adaptive_clbf | .m | combine GP and CBF |
| QPSolver | .m | CBF |
| ModelGP | .m | Interface for gaussian process model |
| ScaledGP | .m | Gaussian process model |
| DynamicsAckermannZ | .m | Approximate dynamics |
| DynamicsAckermannZModified | .m | True dynamics |
| BarrierAckermannPointZ | .m | barriers class |
| LyapunovAckermannZ | .m | CLF |
| circles | .m | outside helper function |
| einsum | .m | outside helper function |