Preface

    What Is the System Identification Toolbox?
    Using This Guide
    Typographical Conventions
    Related Products
    About the Author

The System Identification Problem

    Basic Questions About System Identification
    Common Terms Used in System Identification
    Basic Information About Dynamic Models
        The Signals
        The Basic Dynamic Model
        Variants of Model Descriptions
        How to Interpret the Noise Source
        Terms to Characterize the Model Properties

    The Basic Steps of System Identification
    A Startup Identification Procedure
        Step 1: Looking at the Data
        Step 2: Getting a Feel for the Difficulties
        Step 3: Examining the Difficulties
        Step 4: Fine Tuning Orders and Disturbance Structures
        Multivariable Systems

    Reading More About System Identification

The Graphical User Interface

    The Big Picture
        The Model and Data Boards
        The Working Data
        The Views
        The Validation Data
        The Work Flow
        Management Aspects
        Workspace Variables
        Help Texts

    Handling Data
        Data Representation
        Getting Input-Output Data into the GUI
        Taking a Look at the Data
        Preprocessing Data
        Checklist for Data Handling
        Simulating Data

    Estimating Models
        The Basics
        Direct Estimation of the Impulse Response
        Direct Estimation of the Frequency Response
        Estimation of Parametric Models
        ARX Models
        ARMAX, Output-Error and Box-Jenkins Models
        State-Space Models
        User Defined Model Structures

    Examining Models
        Views and Models
        The Plot Windows
        Frequency Response and Disturbance Spectra
        Transient Response
        Poles and Zeros
        Compare Measured and Model Output
        Residual Analysis
        Text Information
        LTI Viewer
        Further Analysis in the MATLAB Workspace

    Some Further GUI Topics
        Mouse Buttons and Hotkeys
        Troubleshooting in Plots
        Layout Questions and idprefs.mat
        Customized Plots
        What Cannot be Done Using the GUI

Tutorial

    Overview
    The Toolbox Commands
    An Introductory Example to Command Mode
    The System Identification Problem
        Impulse Responses, Frequency Functions, and Spectra
        Polynomial Representation of Transfer Functions
        State-Space Representation of Transfer Functions
        Continuous-Time State-Space Models
        Estimating Impulse Responses
        Estimating Spectra and Frequency Functions
        Estimating Parametric Models
        Subspace Methods for Estimating State-Space Models

    Data Representation and Nonparametric Model Estimation
        Data Representation
        Correlation Analysis
        Spectral Analysis
        More on the Data Representation in iddata

    Parametric Model Estimation
        ARX Models
        AR Models
        General Polynomial Black-Box Models
        State-Space Models
        Optional Variables

    Defining Model Structures
        Polynomial Black-Box Models: The idpoly Model
        Multivariable ARX Models: The idarx Model
        Black-Box State-Space Models: the idss Model
        Structured State-Space Models with Free Parameters: the idss Model
        State-Space Models with Coupled Parameters: the idgrey Model
        State-Space Structures: Initial Values and Numerical Derivatives

    Examining Models
        Parametric Models: idmodel and its children
        Frequency Function Format: the idfrd model
        Graphs of Model Properties
        Transformations to Other Model Representations
        Discrete and Continuous Time Models

    Model Structure Selection and Validation
        Comparing Different Structures
        Impulse Response to Determine Delays
        Checking Pole-Zero Cancellations
        Residual Analysis
        Model Error Models
        Noise-Free Simulations
        Assessing the Model Uncertainty
        Comparing Different Models
        Selecting Model Structures for Multivariable Systems

    Dealing with Data
        Offset Levels
        Outliers and Bad Data; Multi-Experiment Data
        Missing Data
        Filtering Data: Focus
        Feedback in Data
        Delays

    Recursive Parameter Estimation
        The Basic Algorithm
        Choosing an Adaptation Mechanism and Gain
        Available Algorithms
        Segmentation of Data

    Some Special Topics
        Time Series Modeling
        Periodic Inputs
        Connections Between the Control System Toolbox and the System Identification Toolbox
        Memory/Speed Trade-Offs
        Local Minima
        Initial Parameter Values
        Initial State
        The Estimated Parameter Covariance Matrix
        No Covariance
        nk and InputDelay
        Linear Regression Models
        Spectrum Normalization and the Sampling Interval
        Interpretation of the Loss Function
        Enumeration of Estimated Parameters
        Complex-Valued Data
        Strange Results

Functions -- By Category

    Help Functions
    The Graphical User Interface
    Simulation and Prediction
    Data Manipulation
    Nonparametric Estimation
    Parameter Estimation
    Model Structure Creation
    Manipulating Model Structures
    Model Conversions
    Model Analysis
    Model Validation
    Assessing Model Uncertainty
    Model Structure Selection
    Recursive Parameter Estimation
    General

Functions -- Alphabetical List

Examples

Printable Documentation (PDF)

Product Page (Web)