Skip to content
This repository has been archived by the owner on Oct 11, 2024. It is now read-only.

AUVSL/auvsl_dynamics

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

26 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

auvsl_dynamics

This repo contains code related to vehicle dynamics written by Justin Yurkanin ([email protected]) The different branches refer to different experiments. float_model uses Robcogen templated with float scalar type and is fastest and should be used for path planning and anything serious. bekker_descent uses an auto-diff scalar type for doing gradient descent to identify bekker parameters. hybrid_network is similar and trains a physics based neural ODE. 2d_neural_ode is a test branch for training 2D neural ode's for vehicle dynamics. adjoint_method is another test branch to see if the adjoint sensitivity method works better for neural ODE's than just directly calculating the parameter/loss gradient with auto diff.

This repo requires CppAD for auto-diff, libeigen3, and Robcogen for forward dynamics.

For a more in depth description of what all this code does, see my thesis.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages