MATLAB Technology Tour 2011 Agenda
Time
Session Title
08:00
Optional Primer: An Overview of MATLAB and the MATLAB Product Family
08:20
Optional Primer: Introduction to Simulink for Modelling and Simulation
08:50
Welcome and Keynote Presentation: MathWorks from Research to Production
09:30
Streamline Your Data Analysis Workflow with MATLAB
10:00
Data Analysis and Modelling with MATLAB: A Practical Approach
10:30
Morning Tea
11:00
Guest Speaker: Industry Case Study
11:30
Model-Based Design Turns 10
12:00
Pathways to Production Taking MATLAB and Simulink Algorithms from Research and Design to
Production
12:30
Overview of Afternoon Sessions
12:35
Lunch
Workshops (held concurrently)
13:30
Automatically Converting MATLAB Code to C
Code
Data Analysis for Design with MATLAB
14:30
Model-Based Design for High-Integrity and
Business-Critical System Development
Using MATLAB for Signal Analysis
15:40
Closing Remarks, Next Steps, and Q&A
16:00
End
1
Using MATLAB for Signal Analysis
Daryl Ning
Applications Engineer
MathWorks Australia
Level 5, Tower 1
495 Victoria Ave
CHATSWOOD NSW 2067
2011 The MathWorks, Inc.2
Demo: Speech Recognition Application
Technical Computing Workflow
Speech Recognition Application
Access
Explore & Discover
Files
Data Analysis
& Modeling
Software
Algorithm
Development
Share
Reporting and
Documentation
Outputs for Design
Code & Applications
Hardware
Application
Development
Deployment
Automate
4
Agenda
Explore and Discover
Spectral Estimation
Filter Design
Application Specific Tools
Data Access
Data Acquisition
Communicating with Instruments
System Objects for Streaming Applications
Summary
Signal Analysis Examples
Physiological Signals
EEG, ECG, EMG
Geophysical signals
Seismic or tidal data
Communications
SDR, RADAR
Images or Video
Medical, Surveillance, Security
Audio
Speech, Music
Explore and Discover
General Signal Analysis Algorithms
Explore & Discover
Signal transforms
fft, dct, czt, hilbert, fwht,
Signal Analysis
& Visualization
Windowing methods
hamming, hann, bartlett, rectwin,
Algorithm
Development
Spectral estimation algorithms
periodogram, pwelch, msspectrum,
Application
Development
Statistical signal processing
xcorr, xcov, tfestimate, corrmtx,
Multirate operations
interp, decimate,resample,upfirdn,
Explore and Discover
Design and Implement Digital Filters
Explore & Discover
FIR Filters
fir1,fir2,fircls,firpm,intfilt,
Signal Analysis
& Visualization
IIR Filter Design Methods
butter,cheby1,cheby2,ellip,yulewalk,
Algorithm
Development
Filter Analysis and Implementation
fvtool,abs,angle,filter,cost,fftfilt,
Application
Development
Specialized Response Types
fdesign.notch,fdesign.octave,
Adaptive and Multirate Filters
adaptfilt.lms, adaptfilt.rls,
mfilt.firdecim, mfilt.iirdecim,
8
Demo: Remove noise from a Signal
Band-stop Filter
Interactive Tools
SPTOOL
10
Interactive Tools
FDATOOL*
*Some advanced features requires Filter Design HDL Coder, Fixed Point Toolbox and/or DSP System Toolbox.
11
Application Specific Tools
Communications System Toolbox
Image Processing Toolbox
Computer Vision System Toolbox
Phased Array System Toolbox (NEW)
RF Toolbox
Wavelet Toolbox
12
Ford Motor Company Develops and
Deploys Sound-Quality Metrics with
MATLAB
Challenge
Develop and deploy sound-quality metrics that correlate
well with subjective impressions of sound
Solution
Use the MATLAB product family to develop a sound
quality analysis tool and deploy SQ metrics to the
company and its worldwide suppliers
GUI for spark knock detector.
Results
Improved quality of Ford products
Development time reduced by six months
Source code control
Link to user story
13
Argonne National Laboratory Develops
Sensor for Chemical Agent Detection Using
MathWorks Tools
Challenge
Develop an instrument to reliably detect and identify
gaseous chemicals
The Smart Sensor Developer Kit.
Solution
Use MATLAB and related toolboxes to develop a highly
sensitive microsensor that can be incorporated into a
wide range of instruments
A scientist needs to focus on
Results
science and shouldnt have to
Measurement software code performance
increased by two orders of magnitude
Research and development time reduced from one
year to two weeks
SSDK recognized with prestigious award
recreate or debug math libraries
or support code. MathWorks
tools are proven, so researchers
can concentrate on new
algorithms and approaches.
Dr. Michael Vogt
Argonne National Laboratory
14
AT4 wireless Increases Internal Test
Coverage to Over 90% for LTE Physical
Layer Test Equipment Designs
Challenge
Develop test systems for LTE wireless equipment
AT4 wireless LTE layer 1 tester.
Solution
Use MATLAB and Simulink to design and simulate the
LTE physical layer, verify the FPGA implementation,
and analyze test results
Results
Internal test coverage increased to over 90%
Test harness reused throughout the project
life cycle
Development effort reduced by 2530%
MATLAB is a universal language that
makes it easy to exchange algorithms
and test results across our team. Our
physical layer model in MATLAB and
Simulink enabled us to better
understand the LTE specifications, and
Model-Based Design enabled us to
verify that our FPGA implementation
conformed to those specifications.
Francisco Javier Campos
AT4 wireless
Link to user story
15
Institute for Biodiagnostics, Canada
Develops an Imaging Tool for Assessing
Burn Injuries
Challenge
Determine the severity of burn injuries more quickly and
accurately
Images of a first-degree
burn, generated in three
wavelengths from a
spectroscopic camera.
Solution
Infrared point
spectroscopy
probe on the
tip of a finger.
Use MathWorks products to develop a tool to monitor
variations in skin hemodynamics at the site of the wound
Results
Effective, noninvasive tool for accessing burns
Flexible test environment
Accelerated algorithm development process
MathWorks tools allowed us
to thoroughly explore our data,
understand it, and put together
an optimal data analysis
approach.
Dr. Lorenzo Leonardi
Institute for Biognostics, Canada
16
Automating Digital Filter Implementation
for an Audio Codec at MediaTek
Challenge
Design audio codec filters to rigorous signal-to-noise ratio
and total harmonic distortion standards while minimizing
power consumption and silicon area
Filter architecture and segment of generated
code.
Solution
Design individual filters in MATLAB and Filter Design
Toolbox and use Filter Design HDL Coder to generate
synthesizable RTL code
Results
RTL development cycle reduced from three months
to less than two weeks
System modifications made in days, not weeks
Evaluation of multiple filter architectures accelerated
Tests of the initial chip run
uncovered no problems
whatsoever with the digital
portion of the audio codec. This
meant that our team was free to
lend a hand with the analog parts
of the design and with the rest of
the testing effort.
MediaTek Inc.
17
Summary
Explore and Discover
MATLAB speeds the development of signal processing
algorithms with built in tools for
Spectral Analysis
Filter Design
Filter Implementation
Extensive documentation and examples to help you get
started
Easy to use GUIs allow you to quickly explore your
data
A high level programming environment allows you to
perform more advanced and customised analysis
18
Agenda
Explore and Discover
Spectral Estimation
Filter Design: FDATOOL and Filterbuilder
Application Specific Tools
Data Access
Data Acquisition
Communicating with Instruments
System Objects for Streaming Applications
Summary
19
Import, Create or Acquire a Signal
Access
Files
Software
Import signals from multimedia files
(Audio) .wav, .wma, .aif, .aifc, .aiff, .mp3, .au, .sn
(Video) .qt, .mov, .avi, .asf, .asx, .wmv, .mpg, .mp2, .mp4
Generate standard waveforms in MATLAB
Sine, sawtooth, square, chirp, Gaussian pulse,
Code & Applications
Instruments/Devices
Acquire signals from an instrument or device
Signal analyzers, oscilloscopes, data acquisition boards,
audio sound cards,
20
MATLAB Connects to your Hardware Devices
Instrument Control Toolbox
Instruments and RS-232
serial devices
Data Acquisition Toolbox
Plug-in data acquisition devices
and sound cards
Image Acquisition Toolbox
Image capture devices
MATLAB
Interfaces for communicating
with everything
21
Demonstration Setup
GPIB Control Commands
Analog Waveform
Windows Sound
Card (A/D)
Agilent 33120A
Arbitrary Waveform
Generator
22
Demonstration Part 1:
Working with an Agilent Arbitrary Waveform Generator
fwrite(obj, TRIG:MAI:EDGE:SLO RIS);
set(obj, TriggerSlope, Rising);
23
Demonstration Part 2:
Acquiring Data with the Data Acquisition Toolbox
24
Instrument Control Toolbox
Control and communicate with scientific instruments
Integrate instruments into
MATLAB applications and
Simulink models
Interactive tool for detecting and
controlling instruments
Automatic code generation for
faster and easier implementation
Support for IVI, VXIplug&play,
and MATLAB instrument drivers
Support for common
communication protocols
25
Instrument Control Toolbox
Communication Protocols
GPIB Boards
Agilent Technologies
CONTEC
Keithley
National Instruments
Capital Equipment Corporation (CEC)
IOTech
Measurement Computing
ICS Electronics
VISA Interface (includes Serial, GPIB, VXI, GPIB-VXI, TCP/IP, USB)
Agilent
Rohde & Schwarz
National Instruments
Tektronix
Network Protocols
TCP/IP and UDP
Serial Port (core functionality in MATLAB)
RS-232, RS-422, RS-485
Expanded capability with Instrument Control Toolbox
For a full listing, visit: www.mathworks.com/products/instrument
26
Instrument Control Toolbox
Instrument Drivers
Industry-standard instrument drivers
LAN eXtensions for Instrumentation (LXI)
Interchangeable Virtual Instrument (IVI) drivers
VXIplug&play drivers
Several hundred supported instruments
Drivers available through the instrument vendor
Instrument drivers do not require knowledge of lowlevel commands
fwrite(obj, TRIG:MAI:EDGE:SLO RIS);
set(obj, TriggerSlope, Rising);
27
Instrument Control Toolbox
Instrument Drivers
MATLAB Instrument
Driver
Vendor Instrument
Driver
(IVI or VXIplug&play)
Automatically create MATLAB instrument driver for vendors
driver
Customize MATLAB driver to extend or enhance vendors driver
Create custom MATLAB driver for any instrument
MATLAB instrument drivers available at:
www.mathworks.com/products/instrument/drivers
28
Data Acquisition Toolbox
Acquire and output data from data acquisition boards
Immediately analyze live or
acquired data in MATLAB and
Simulink
Configure hardware without
leaving MATLAB
Incorporate custom analysis into
PC-based digital oscilloscope
Ability to do one-shot or
continuous acquisition
Support multiple data acquisition
devices and vendors
29
Data Acquisition Toolbox
Supported Hardware
Acqiris*
ADLINK*
Agilent*
Advantech
CONTEC*
Data Translation*
g.tec*
IOTech*
Keithley
Measurement Computing (MCC)
National Instruments
Ono Sokki*
United Electronic Industries*
VXI Technology
Any PC compatible sound card (AI, AO)
Parallel Port (DIO)
* Denotes that the hardware manufacturer made the investment to provide this support
For a full support listing, visit: www.mathworks.com.au/products/daq
30
Image Acquisition Toolbox
Acquire images and video from industry-standard hardware
Stream images and video
directly into MATLAB
Configure device properties
Background image acquisition
Live video previewing
Support for multiple hardware
vendors
31
Image Acquisition Toolbox
Supported Hardware
Frame grabbers
Matrox Imaging
Data Translation
DALSA Coreco
National Instruments
Digital cameras
Hamamatsu
Qimaging
DCAM-compatible cameras (IIDC 1394-based digital camera
specification)
Windows video devices
Windows Driver Model (WDM) or Video For Windows (VFW) drivers
Full listing available at www.mathworks.com/products/imaq
32
OPC Toolbox
Connect MATLAB And Simulink applications to OPC servers.
The OPCTool GUI enables easy
browsing of a server namespace,
configuring objects, and reading
and writing data.
The OPC Blockset library lets
you communicate data directly
to and from Simulink and your
process system.
33
Vehicle Network Toolbox
Send/Receive CAN packets directly from MATLAB or Simulink
Functions for Tx/Rx of CAN packets
Functions for encoding, decoding,
filtering and logging of CAN messages
Communication blocks for connecting
a CAN bus to a Simulink model
Message traffic GUI for visualising live
CAN network traffic
Support for Vector interface hardware
and Vector CAN database (.dbc) files
34
Newport Corporation uses MATLAB to validate positional
accuracy of motion control devices used for microchip
fabrication
Because we use the
Instrument Control and Data
Acquisition Toolboxes, all the
Software used:
MATLAB
Data Acquisition
Toolbox
Instrument Control
Toolbox
test data is in MATLAB, and
I can obtain the results
immediately . MathWorks
tools saved us hundreds of
hours of testing time.
Reed Farrar
Newport Corporation
35
Summary
Data Access
MATLAB connects to hardware and instruments
Control instruments directly from MATLAB
Acquire live signals directly into MATLAB
Data access and exploration can be integrated into one
environment
No need to pipeline data between different software
applications
A more streamlined workflow
36
Agenda
Explore and Discover
Spectral Estimation
Filter Design: FDATOOL and Filterbuilder
Application Specific Tools
Data Access
Data Acquisition
Communicating with Instruments
System Objects for Streaming Applications
Summary
37
Introducing System Objects
System objects are MATLAB objects for time-based and
data-driven algorithms, sources, and sinks
System objects make stream processing in MATLAB much
easier
Advantages of system objects over MATLAB functions:
Automatically manage state information
Manage data indexing
Implicit data buffering
38
Batch Processing
Load the entire data set and process it all at once
MATLAB
Data
Files
Memory
Signal
Processing
Algorithm
39
Batch Processing
Loads entire dataset
into workspace
[audio, Fs] = wavread('dspafxf_8000.wav');
filtCoeff = fir1(40, 0.8,low');
audio_filt = filter(filtCoeff, 1, audio);
wavplay(audio_filt, Fs);
Requires entire dataset to play audio
40
Stream Processing
Load a frame of the signal and process it before moving
on to the next frame
MATLAB
Stream
Source
Memory
Stream
Processing
41
Traditional Stream Processing in MATLAB
%% Streaming the MATLAB way
% set up initializations
filename = 'dspafxf_8000.wav';
Fs = 8000;
info = mmfileinfo(filename);
num_samples = info.Duration*Fs;
frame_size = 40;
filtCoeff = fir1(40, 0.8,low');
filtState = zeros(1,numel(filtCoeff)-1);
output = zeros(1,num_samples);
More code than in batch
processing example
Explicit state management
Explicit indexing
%% Processing in the loop
index= 1;
while index < (num_samples-frame_size+1)
data = wavread(filename,[index index+frame_size-1]);
[datafilt,filtState] = filter(filtCoeff,1,data,filtState);
output(index:index+frame_size-1) = datafilt;
index = index + frame_size;
end
wavplay(output,Fs);
Need to maintain output buffer
42
System Objects Make Stream Processing
Easier
Initialize objects
% set up initializations
filename = 'dspafxf_8000.wav';
hAudio = dsp.AudioFileReader(filename, ...
'SamplesPerFrame',40,OutputDataType','double');
hFilt = dsp.DigitalFilter(
TransferFunction, 'FIR (all zeros),
Numerator, fir1(40, 0.8,low');
hPlayer = dsp.AudioPlayer('SampleRate', 8000);
%% Processing in the loop
while ~isDone(hAudio)
data
= step(hAudio);
datafilt = step(hFilt, data);
step(hPlayer, datafilt);
end
In-the-loop code is much simpler
Implicit states and indexing
Audio player runs in-the-loop
43
Demo: Noise Cancellation with an Adaptive
LMS Filter
Noisy signal
d(n)
s(n) + z(n)
+
Z(n)
Correlated
to noise
Adaptive
filter
Adapting
algorithm
y(n)
Clean
signal
e(n)
d(n) y(n)
For details, see "Adaptive Filter Theory", 3rd Ed. by S. Haykin, Prentice-Hall 1996.
44
Working with System Objects
Examples from the DSP Systems Toolbox
Signal processing System objects
dsp.<objectname>
Eg. Digital filter System object
dsp.DigitalFilter
Create a System object with
default parameters
hFilt = dsp.DigitalFilter
Query object parameters and
methods
get(hFilt, Property Name)
methods(hFilt)
Example method step:
data_out = step(hFilt, data_in)
Use the step method to execute the
algorithm in a loop
45
Where Can I Find System Objects?
DSP System Toolbox
Previously the Signal Processing Blockset
Communications System Toolbox
Previously the Communications Toolbox and Communications
Blockset (now combined into a single toolbox)
Computer Vision System Toolbox
Previously the Video and Image Processing Blockset
Phased Array System Toolbox
Brand new to R2011a!
46
Summary
System Objects for Streaming Applications
System Objects make stream processing in MATLAB
much easier
System Objects retain state information
No need to manually track indices and maintain buffers
MATLAB in the loop code becomes simpler
You dont need to be an expert in object oriented
programming to use System Objects
47
Agenda
Explore and Discover
Spectral Estimation
Filter Design: FDATOOL and Filterbuilder
Application Specific Tools
Data Access
Data Acquisition
Communicating with Instruments
System Objects for Streaming Applications
Summary
48
Technical Computing Workflow
Access
Explore & Discover
Files
Data Analysis
& Modeling
Software
Algorithm
Development
Share
Reporting and
Documentation
Outputs for Design
Code & Applications
Hardware
Application
Development
Deployment
Automate
49
Summary
Using MATLAB for Signal Analysis
MATLAB is a powerful environment for design and simulation
of signal processing algorithms
Signal analysis and visualization tools
Signal processing operations and algorithms
Digital filter design and implementation
Interfaces with hardware and instruments
Stream processing techniques with system objects
Application specific tools for Communications, Image/Video
Processing, RF, and Phased Array systems
50
Further Information
DSP Application Page:
http://www.mathworks.com.au/dsp/
Technical Literature:
http://www.mathworks.com.au/dsp/technicalliterature.html
MathWorks Australia Pty Ltd
Level 5, Tower 1
Email: [email protected]
Phone: 02 8669 4700
495 Victoria Ave
CHATSWOOD NSW 2067
Fax: 02 8669 4700
Web: www.mathworks.com.au
51
Upcoming Training in Melbourne
http://www.mathworks.com.au/services/training/schedule.html
Jun
22 24 MATLAB Fundamentals
27
MATLAB for Data Processing and Visualisation
28
MATLAB Programming Techniques
29
MATLAB for Building Graphical User Interfaces
Aug
2 4 MATLAB Fundamentals
18 19 Signal Processing with MATLAB
22 23 Simulink for System and Algorithm Modelling
Interested in an on-site course? Let us know.
52
Upcoming Training in Sydney
http://www.mathworks.com.au/services/training/schedule.html
Jun
21 23 MATLAB Fundamentals for Financial Applications
28 29 Parallel Computing with MATLAB
Jul
20
Interfacing MATLAB with C Code
26 28 MATLAB Fundamentals
Aug
15 16 Signal Processing with MATLAB
Interested in an on-site course? Let us know.
53
Upcoming Training in Brisbane
http://www.mathworks.com.au/services/training/schedule.html
Sep
20
21
22
MATLAB for Data Processing and Visualisation
MATLAB Programming Techniques
MATLAB for Building Graphical User Interfaces
Oct
25 27 MATLAB Fundamentals
Interested in an on-site course? Let us know.
54
Upcoming Training in Adelaide
http://www.mathworks.com.au/services/training/schedule.html
Jun
15 17 MATLAB Fundamentals
Aug
9 10 Simulink for System and Algorithm Modelling
Sep
68
MATLAB Fundamentals
Interested in an on-site course? Let us know.
55
Upcoming Training in Perth
http://www.mathworks.com.au/services/training/schedule.html
Aug
30
31
MATLAB for Data Processing and Visualisation
MATLAB Programming Techniques
Sep
MATLAB for Building Graphical User Interfaces
Interested in an on-site course? Let us know.
56
