System Identification Toolbox Getting Started Introduction What Is the System Identification Toolbox? 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 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 Simple Process Model Estimation of Parametric Models ARX Models ARMAX, Output-Error (OE), and Box-Jenkins (BJ) 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 Outputs 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 Examples Tutorial Overview The Toolbox Commands An Introductory Example to Command Mode Example Details 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 The advice Command Data Representation and Nonparametric Model Estimation Data Representation Correlation Analysis Spectral Analysis Frequency Domain Data More on the Data Representation in iddata Parametric Model Estimation ARX Models AR Models General Polynomial Black-Box Models Process Models State-Space Models Optional Variables Defining Model Structures Polynomial Black-Box Models: the idpoly Model Process Models: the idproc 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 Estimating Continuous-Time Models: General Remarks 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; Multiple-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 Tradeoffs Local Minima Initial Parameter Values Initial State Initial States for Frequency Domain Data 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 -- Categorical List 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 Release Notes System Identification Toolbox 6.1.1 Release Notes Major Bug Fixes System Identification Toolbox 6.0 Release Notes New Features New Model Object Estimation and Validation in the Frequency Domain Storing Continuous-Time Data as Frequency Domain Objects Simulink Support of iddata and idmodel Objects Advice Command Upgrading from an Earlier Release Theta Models are no Longer Supported System Identification Toolbox 5.0 Release Notes New Features Object-Based Design Advanced Feature Enhancements Printable Documentation (PDF) Product Page (Web)