Recursive Least Squares (RLS) Identification in MATAB

An algorithm for Recursive Least Squares (RLS) model parameters identification using MATLAB.

This project provides a MATLAB implementation of the Recursive Least Squares (RLS) algorithm for identifying system parameters. The algorithm is designed to estimate the parameters of a discrete-time transfer function model based on input and output data.

Features

• Parameter Identification: Identify the system parameters for a given system input and output using the RLS method.
• Data Preprocessing: Preprocess data based on step changes of the manipulated variable and conduct centralization and normalization.
• Plotting: Visualize the convergence of parameters and compare the actual and estimated system outputs.

Installation

Via tbxmanager (recommended)

tbxmanager install rls-identification

See tbxmanager.com if you don't have tbxmanager yet.

From source

1. Clone the repository:

   git clone https://github.com/MarekWadinger/RLS_identification.git

2. Navigate to the project directory:

   cd RLS_identification

Usage

1. Load your data into MATLAB. The data should be in a structure with fields t (time), u (input), and y (output).

2. Preprocess the data:

   [u_mean, y_mean, idx] = preprocessData(data);

3. Identify the system parameters:

   Ts = 0.01;
   idtf = recursiveLeastSquares(u_mean, y_mean, Ts, 1, 2, 'PlotConv', true);

Example

Here is an example of how to use the provided functions:

%% Load file
load("example_fan_control.mat");

%% Make preprocessing
[u_mean, y_mean, idx] = preprocessData(data);

%% Identify system parameters of the specified order 
Ts = 0.01;
idtf = recursiveLeastSquares(u_mean, y_mean, Ts, 1, 2, 'PlotConv', true);

License

MIT License