1
Model-Based Design for Signal Processing and Communications: Automotive Applications
Dr. Colin Warwick, The MathWorks, Inc.
© 2004 The MathWorks, Inc.
MATLAB, Simulink, Stateflow, and Real-Time Workshop are registered trademarks of The MathWorks, Inc.
Other product or brand names are trademarks or registered trademarks of their respective holders.
2
Agenda
Advancing Technologies in the Automobile– In which do signal processing and communications
have a role?Model-Based Design Addresses Critical Challenges– Leverage control design methodology into project
that use signal processing and communication– Benefits
Customer SuccessesSimulink 6 family of products drives the evolution and adoption of Model-Based Design
2
3
Software and electronics are pervasive in in today’s vehicles
Power Management
Transmission
EngineRide Control
ABS
Steering
Stability Controls
Traction Control
Obstacle Detection
Adaptive Cruise Control
Crash Avoidance
Airbags
Adaptive Front Lighting Systems
Passenger Detection
WindowsDoors
Lights
Climate Controls
Driver Drowsiness Infotainment
Instrumentation
Voice Recognition
Navigation
Wireless Connectivity
4
ChallengesExample: Wireless communications
Time-to-market pressureIncreasing complexity– Multiple standards– Voice, video, and data
integration– Higher data rates
Broad palette of target technologies– Team integration– Moving partitioning
boundaries
ASSPAnalog
FPGA
ASIC DSP s/w
MCU s/w
JTAG!
IP3!ACS!
BIST!
RTOS!
IRQ!
3
5
SpecDesign
ImplementTest
Introduced
Detected
8%
15%
22%
55%
60%
21%
12%
7%0%
10%
20%
30%
40%
50%
60%
70%
Where Errors Are Introduced... and Detected
Source: “Migration from Simulation to Verification with ModelSim”by Paul Yanik. EDA Tech Forum, 2004 Mar 11, Newton MA
Close the verification gap
“…each delay in the detection and correction of a design problem makes it an order of magnitude more expensive to fix…”
Clive Maxfield and Kuhoo Goyal “EDA: Where Electronics Begins”
TechBites Interactive, October 1, 2001ISBN: 0971406308]
6
ChallengesExample: Wireless communications
Time-to-market pressureIncreasing complexity– Multiple standards– Voice, video, and data
integration– Higher data rates
Broad palette of target technologies– Team integration– Moving partitioning
boundaries
ASSPAnalog
FPGA
ASIC DSP s/w
MCU s/w
JTAG!
IP3!ACS!
BIST!
RTOS!
IRQ!
4
7
The Value of Model-Based DesignModel-Based Design
Executable specificationDesign with simulationImplementation through code generationContinuous test and verification
Accelerates InnovationRapid design iterations“What-if” studiesUnique features and differentiators
Improves Quality Minimize hand coding errorsUnambiguous communication internally and externally
Reduces Development CostReduce expensive physical prototypesEliminate errors early when they are cheap to fixReduce testing
Improves Time-to-marketGet the design right the first time
8
Model-Based Design applied to Signal Processing and Communications
Simulink
MATLAB
Toolboxes
Sign
al DAQ CardsInstruments
Link for Code ComposerLink for ModelSim
Desktop Apps.Automated Reports
Blocksets
Sign
al
5
9
Frames Simplify Multi-rate Models
ts=2
ts = 3
Execution sequence
Sample clock, ts = gcd(2,3) = 1
Sample-by-sampleprocessing
without frames
Process 1
Process 2
Frame clock, tf = gcd(6, 6) = 6
Execution sequence
tf=6With framesA frame is
a sequence of samplesgrouped together for execution
Note: An important side benefit of frames is execution speed
10
Nallatech Designs a Reconfigurable Video Encryption System for the UK Ministry of Defence with Simulink
The ChallengeTo design a reconfigurable video encryption system
The SolutionUse Simulink and Xilinx System Generator for DSP to create algorithms represented in subsystems, debug and integrate the core, format the video, and generate VHDL code for the FPGA
The ResultsDesign time reduced by 2/3VHDL training avoidedCollaborative development accomplished remotely
“We could not have implemented the algorithms on the FPGA without Simulink and Xilinx System Generator."
Malachy Devlin,Nallatech
Reconfigurable video encryption system.
www.mathworks.com/user_story
6
11
Zyray Wireless develops low-cost semiconductor product with MathWorks tools
The ChallengeTo develop a low-cost semiconductor product based on 3G standards that handset manufacturers could combine with chips based on 2G standards
The SolutionUse MathWorks tools to develop algorithms and model the chip subsystems
The ResultsChip saves manufacturers millions of dollarsModels reused for production releaseDevelopment time cut in half
“MATLAB is an ideal environment for developing and understanding our algorithms. Simulink integrates well with MATLAB and lets us produce a design that looks very similar to what we end up with ultimately in hardware."
Francis SwartsZyray Wireless
SPINNERchip add-on WCDMA baseband processor
www.mathworks.com/user_story
12
Yet, engineers are asking for…
Modeling of ever larger systemsManaging multi-team development projectsBetter integration with coding practicesExceed hand code performanceDesign more types of systems and components, including hardware Tighter integration of Simulink & MATLAB
These are the next set of challenges
7
13
Simulink 6
Large-scale projects and end-to-end development
New domains and applications
…driving the evolution and adoption of Model-Based Design
14
Simulink Family of ProductsSimulinkSignal Processing BlocksetCommunications BlocksetRF BlocksetFixed-point ToolboxSimulink Fixed PointFilter Design Toolbox– Filter Design HDL Coder
Link for ModelSimVideo and Image Processing BlocksetFamily of 45* products in total
*includes the 13 xPC TargetBox hardware products
8
15
Simulink 6 for Large-Scale Projectsand End-to-End Development
Large-scale modeling and teams– Model Reference– Model Explorer
End-to-end embeddedsystem development– Embedded MATLAB Function Block – Fixed-point math – Production code– Test and verification
16
From textbook to Model-Based Design
uwHny =
nn yde −=
∆uu∆uG H+
=λ
)(1 ∆Gu∆∆ H−←λ
*eGww +←
Output:
Error:
Gain vector(Mx1):
Inv. corr. matrix(MxM):
Weight update(Mx1):
9
17
Demonstrations of Simulink 6 Family for New Domains and Applications
IEEE 802.11a wireless LAN model– RF Blockset– Embedded MATLAB Function
BlockImage stabilization– Video and Image Processing
Blockset
18
DEMO: WLAN Communications Model
Description– End-to-end model of 802.11a
Wireless LAN physical layer
What you will see– One model for baseband, RF,
and channel– Fast time-domain RF simulation
with RF Blocksets– Embedded MATLAB function block
10
19
WLAN Communications ModelEmbedded MATLAB Function Block
Blocks from new RF Blockset
Integration with the new RF Toolbox
20
DEMO: Video Motion Stabilization
Description– Closed-loop embedded
video processing
What you will see– Video and Image
Processing Blockset– Code generation for TI
C6000 DSP
11
21
Video Motion Stabilization
Double, single, integer
Code generation for TI C6000 DSP
New Video and Image Processing Blockset
Streaming video in Simulink
22
Generate VHDL and Verilog Code for Fixed-point Filters from MATLAB
Filter Design HDL CoderFixed-point filters from Filter Design ToolboxEfficient, synthesizable, and portable VHDL or Verilog codeASIC or FPGA targetsAutomatic creation of VHDL, Verilog, and ModelSim test benches
12
23
Co-simulate with ModelSim®
using Link for ModelSim
Links ModelSim PE or SE with MATLAB and SimulinkVHDLVerilog or SystemC with VHDL wrappers
Simulink ModelSim
24*as of 2003Q3
365* add-on products and services from partners that complement and extend MathWorks productsSpecialized third-party toolboxes for MATLABInterfaces to partners’ software and hardware productsSpecialized training courses and consulting servicesSystem integrators and suppliers that incorporate MathWorks’ products
13
25
The MathWorks’ Family of Productswith Xilinx System Generator for DSP
Simulink
MATLAB
Stateflow
Blocksets
ANSI C andIntrinsics forMCUs & DSPs
Coders & Targets
Real-timeWorkshop
See your local MathWorks’ sales person for details on application-specific products
Toolboxes
HDL forXilinx FPGAs
DAQ CardsInstruments
Link for Code ComposerLink for ModelSim
Desktop Apps.Automated Reports
26
SummarySimulink 6 brings:– Model-Based Design to large-scale
projects– More comprehensive coverage of
embedded system development – New domains and applications
Simulink continues to drive the evolution and adoption of Model-Based DesignVisit the web for morewww.mathworks.com/r14