145 COMP UNC-Chapel Hill Maintenance Manual I: An Annotated Outline The Kalman Filter Online Learning Tool
Greg Welch, client
March 8, 2001
Chris Riley, Director Tom Bodenheimer, Producer
Erin Parker, Admin. Leader John Carpenter, Librarian
1 DOCUMENT CHANGE HISTORY
1
1 Document Change History
March 8, 2001
Annotated Outline
Contents
1 Document Change History 2 Introduction 2.1 Statement of purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Document conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Other relevant documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 System Overview 3.1 High-level view of the system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Directory structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Procedures 4.1 Editing the source . . . . . . 4.2 Building the system . . . . . 4.3 Testing the system . . . . . . 4.4 Updating the documentation 1 2 2 2 2 3 3 4 4 5 5 5 5 5 6 6 6 6 7 8
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
A The User Manual A.1 Quick Start . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.2 Descriptions of the parameter settings and control buttons . . . . . . . . . . . . . . . . . . . A.3 Stepping mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B Development System Requirements and Information C Contact Information
2 INTRODUCTION
2
2 Introduction
2.1 Statement of purpose
The purpose of this document is to provide instructions and guidance for the client (or technical staff appointed by the client) to use when extending/modifying The Kalman Filter Online Learning Tool. The Kalman lter is a set of mathematical equations that provides an efcient computational estimate of the state of a process (e.g., the position and orientation of an airplane) given a time-varying sequence of noisy measurements (e.g., air speed, pressure, temperature, engine thrust). The lter is a popular mathematical estimator due to its efciency and robustness. The Kalman Filter Online Learning Tool is a web-based tool to help develop the intuition and insight of novice users regarding the behavior of the Kalman lter. Users have the ability to change various input parameters and then see how the Kalman lter responds for a given set of noisy measurements. It is assumed that the reader of this document has a general understanding of The Kalman Filter Online Learning Tool. In particular, the reader should be familiar with how the tool works, having used the tool and read the User Manual (see appendix A). It is also advised that the reader have experience with the JavaTM programming language.
2.2 Document conventions
Source code (function names and parameters) will appear in the typewriter font. Other conventions will be introduced here as they are needed.
2.3 Other relevant documents
If the team decides that including some of the other documents (Contract, Design Specication, etc.) is appropriate, they will be mentioned here and attached as an appendix.
3 SYSTEM OVERVIEW
3
3 System Overview
3.1 High-level view of the system
The Kalman Filter Online Learning Tool is divided into four modules (see gure 1), each performing a distinct function. (The high-level diagram may not look exactly like gure 1, but it will be included here.)
Plotter
Kalman Filter Engine User Interface
Data Repository
Figure 1: High-level view of the The Kalman Filter Online Learning Tool system.
User interface. Primarily, the user interface serves to provide the user with choices that govern the execution of the Kalman lter for a specic test problem. The user selects among several options for the actual dynamics, the estimate, the measurement model and the process noise. The user also selects when to run!in UI and step!in UI through the simulation. Secondarily, the user interface acts as the main controller for the system. It passes parameters to the Kalman lter engine and collects the truth, estimate, covariance and residual values for each time-step of the simulation. The user interface passes such simulation values to the plotter and is responsible for displaying the plots produced. The user interface also passes the values for an entire simulation to the data repository calls and on it for a single time-steps values when in step mode. Kalman lter engine. The Kalman lter engine implements the Kalman lter equations for all models of actual dynamics, estimate and measurement. Plotter. The plotting tool generates plots of the simulation values passed from the user interface. The data repository stores a record of the Kalman lter iterations and values computed.
Data Repository.
3 SYSTEM OVERVIEW
4
3.2 Directory structure
Material for this section will evolve as implementation continues. It will particularly be affected by the installation of revision control software. This section will discuss the organization of project les into directories according to modules, etc.
3.3 Files
This section also has yet to be resolved. It will include a list of all project les, a brief description of their contents and their location.
4 PROCEDURES
5
4 Procedures
4.1 Editing the source
This section will involve an explanation of the how the source code is organized in CVS, as well as, the location of the code. Instructions on using CVS to do the following will be provided. checking out the source code
editing the source code checking in the source code comment procedures for each revision
4.2 Building the system
This section will layout guidelines for building the executable. In particular, the following will be discussed. which departmental machines should be used for this purpose JDK: which and how to use
4.3 Testing the system
This section will make suggestions on how to best test the application. The clients Matlab code and results are a good way to test the accuracy of the data and plots.
4.4 Updating the documentation
The Implementation Manual and Maintenance Manual should be updated when necessary. This section will A provide details on where to get the documents and what format they are written in (LTEX or Microsoft WordTM ).
A THE USER MANUAL
6
A
The User Manual
The User Manual is provided to give the reader some context for how the The Kalman Filter Online Learning Tool is to be used, as well as, to document a set of instructions (should they ever be needed).
A.1 Quick Start
1. Go to http://whatever. 2. Click on Use the The Kalman Filter Online Learning Tool. 3. Leave the controls in their default settings. 4. Click on the run button. 5. View the plots. 6. Close the plot window.
A.2 Descriptions of the parameter settings and control buttons
This section will give short description of each (including options, defaults, actions, etc.) truth estimate measurement process noise run step
A.3 Stepping mode
The stepping mode (possibly being the only complicated feature of the tool) will be carefully explained.
B DEVELOPMENT SYSTEM REQUIREMENTS AND INFORMATION
7
B Development System Requirements and Information
The following are requirements for extensions/mod...