index.md 1.27 KB
Newer Older
hazrmard's avatar
hazrmard committed
1
2
3
4
5
---
title: Documentation - Home
order: -1
---

6
7
8
9
10
11
12
13
14
15
16
17
18
19
# Data Driven Models

This project explores how combinations of physical and data based models fare against actual systems.

Physical models require a rigorous understanding of the system. Even then, they are not equipped to account for noise, degradation, and higher order effects.

Data driven models do not depend on an underlying model of the system. They simply develop mappings between input and output variables. However, by doing this, they forego *a priori* knowledge of the distribution of measurements. That is, they have to build an understanding of the system from scratch which requires more data.

## Appproaches

* Train the data model on a simple physics model to initialize parameters. Then train the data model on the actual readings.

* Make a composite additive model based on physical and machine learning sub-models.

hazrmard's avatar
hazrmard committed
20
21
22
23
24
25
26
## Requirements

* MatLab/Simulink
* System identification toolbox
* Communication systems toolbox
* SimScape library

27
28
29
30
## Project Structure

### `PhysicalModels`

hazrmard's avatar
hazrmard committed
31
32
33
34
35
36
37
This directory contains files that simulate physical systems and return input/output readings. MatLab Simulink/Simscape are used for this purpose.

## Contents

* [Background](0_background.md)

* [Model: Resistor-Capacitor Circuit](1_RL_Circuit.md)