Getting Started

    Preface
        Using This Guide
        Typographical Conventions
        Related Products
        About the Author

    The System Identification Problem
        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

Using the System Identification Toolbox

    The Graphical User Interface
        The Model and Data Boards
        The Working Data
        The Views
        The Validation Data
        The Work Flow
        Management Aspects
        Workspace Variables
        Help Texts
        Handling Data
            Getting Input-Output Data into the GUI
            Taking a Look at the Data
            Preprocessing Data
            Checklist for Data Handling
            Simulating Data

        Estimating Models
            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
            Troubleshooting in Plots
            Layout Questions and idprefs.mat
            Customized Plots
            What Cannot be Done Using the GUI

    Tutorial
        The Toolbox Commands
        An Introductory Example to Command Mode
        The System Identification Problem
            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
            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
            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
            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

Reference

    Command Reference
        Functions Listed Alphabetically
            aic
            Algorithm Properties
            ar
            armax
            arx
            arxdata
            arxstruc
            bj
            bode
            compare
            covf
            cra
            c2d
            detrend
            d2c
            EstimationInfo
            etfe
            ffplot
            freqresp
            fpe
            get
            idarx
            iddata
            ident
            idfilt
            idfrd
            idgrey
            idinput
            idmodel
            idmodred
            idpoly
            idss
            impulse
            init
            ivar
            ivstruc
            ivx
            iv4
            LTI commands
            merge (iddata)
            merge (idmodel)
            midprefs
            misdata
            nkshift
            noisecnv
            nuderst
            nyquist
            n4sid
            oe
            pe
            pem
            plot (iddata)
            plot (idmodel)
            polydata
            predict
            present
            pzmap
            rarmax
            rarx
            rbj
            resample
            resid
            roe
            rpem
            rplr
            segment
            selstruc
            set
            setpname
            sim
            simsd
            size
            spa
            ss, tf, zpk, frd
            ssdata
            step
            struc
            timestamp
            tfdata
            view
            zpkdata

Printable Documentation (PDF)

Product Page (Web)