Skip to content

Physically-informed model discovery of systems with nonlinear, rational terms using the SINDy-PI method. Contains functionality for spectral filtering/differentiation.

License

Notifications You must be signed in to change notification settings

BystrickyK/SINDy

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

58 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SINDy

Discovering the governing equations of nonlinear systems purely from data using the SINDy method (https://www.pnas.org/content/113/15/3932).

Lorenz system simulation with feedforward strongly nonlinear inputs. This simulation data was used for identification. https://user-images.githubusercontent.com/55796835/127328774-a1342b36-4729-4225-8e3e-fd447fc66a5e.mp4

Lorenz system simulation with feedback inputs. Used for demonstration of identification of a controlled system. Note that the state gravitates more strongly to the right attractor, where it's being pushed by the control law. https://user-images.githubusercontent.com/55796835/127328914-39fbc4ed-b294-48a5-9ac7-729e497d5644.mp4

Simulation of the reference Lorenz model and the model identified from its measurement data. https://user-images.githubusercontent.com/55796835/127679128-65357765-4e09-4b51-9d53-e436224f4084.mp4

Simulation of the pendulum-cart system https://user-images.githubusercontent.com/55796835/127646378-228cc491-b344-4ab0-9caf-d83239d9fb81.mp4

Simulation of the reference pendulum-cart model and model identified using SINDy from noisy data. https://user-images.githubusercontent.com/55796835/127749887-452fa243-7efa-4761-86bd-0d9dca009ddb.mp4

Reference model vs Bootstrapped model https://user-images.githubusercontent.com/55796835/127749898-489bcebc-fb2e-4f7f-bea3-ada7994ee317.mp4

SINDy model vs Bootstrapped model https://user-images.githubusercontent.com/55796835/127749894-e9e21523-7ae3-4ad0-b916-287db34bbfc4.mp4

Lorenz reference vs identified, nonlinear and discontinuous input sgn(u1) to dx1 https://user-images.githubusercontent.com/55796835/127779031-6db32467-f23b-4574-8389-3a0982e1bd85.mp4

Model identified from real measurement data https://user-images.githubusercontent.com/55796835/127922658-35e7b487-f04e-431a-a6e0-aed10cce4e76.mp4

The real measurement data, visualized https://user-images.githubusercontent.com/55796835/127937959-67bdb377-eedd-4c89-b2cd-edfd253517ce.mp4