Insert image here
 
TECHNICAL UPDATES
 
Products Updates
 

LATEST Release 2007a is now available!
In March 2007, The MathWorks will ship Release 2007a. The latest release includes updates to MATLAB and Simulink, two new products released since R2007a, and updates and bug fixes to 82 other products.

R2007a adds support for the Intel®-based Mac, Windows Vista™, and 64-bit Sun Solaris™ SPARC platforms. Support for the Solaris 32 platform and all Mac operating systems prior to 10.4.7 will be discontinued.

Among the new capabilities of MATLAB R2007a is performance speed up utilizing multithreaded computation in key MATLAB math functions for multicore and multiprocessor systems.


New capabilities for the Simulink product family include:

  • Multithreaded computation in key MATLAB math functions for multicore and multiprocessor systems
  • Ability to run four parallel MATLAB tasks on your desktop with Distributed Computing Toolbox
  • New categorical and dataset array in Statistics Toolbox
  • Fixed-Point Toolbox acceleration at the speed of compiled C
  • Exact modeling and analysis of control loops with delays in Control System Toolbox
  • Generation of nonlinear models with System Identification Toolbox
  • Simulated annealing in Genetic Algorithms and Direct Search Toolbox

New capabilities for the MATLAB product family include:

  • Multidimensional signal support for Simulink, Signal Processing Blockset, Embedded MATLAB Function block, and Real-Time Workshop
  • Improved code efficiency and MISRA C support in Real-Time Workshop Embedded Coder
  • New graphical interface for analyzing and scaling fixed-point systems in Simulink Fixed Point
  • Vector and matrix support in SimEvents
  • Simscape, a new product for multidomain physical modeling

Major Update Products included in the prerelease

  • Control System Toolbox 8
  • Datafeed Toolbox 2
  • Filter Design HDL Coder 2
  • Financial Derivatives Toolbox 5
  • Fixed-Point Toolbox 2
  • Link for Code Composer Studio 3
  • Link for Cadence® Incisive® 2
  • RF Blockset 2
  • SimEvents 2
  • Statistics Toolbox 6
  • System Identification Toolbox 7
  • Wavelet Toolbox 4


New Products Previously Released via the Web (R2006b+)

  • Link for Cadence® Incisive®

    Link for Cadence® Incisive® is a cosimulation interface that integrates MATLAB and Simulink into the hardware design flow for application-specific integrated circuit (ASIC) and field programmable gate array (FPGA) development.

    Link for Cadence Incisive enables you to verify your HDL design from within MATLAB and Simulink. It provides native cosimulation support for Verilog and support for VHDL and mixed-language simulations via Verilog modules.

    http://www.mathworks.com/products/incisive/


New Products for R2007a - Simscape 1.0

Simscape is the new physical modeling platform for physical modeling tools. It allows modeling of multiple domains in one environment. Simscape contains basic elements for creating custom models of physical systems.

Physical modeling products SimDriveline, SimHydraulics and SimMechanics will require and be based on Simscape.
 
 
Tips and Techniques
 

Debugging MATLAB M-files from the MATLAB Command Prompt

What Debugging Tools Are Available at the MATLAB Command Prompt?
This section describes how to make use of functions to debug programs from the MATLAB command prompt. There are altogether seven topics and in the e-newsletter issue, we will look at the third topic. The rest of the topics will be covered in subsequent e-newsletter issues.

Topics
1. Setting, Clearing, and Querying Breakpoints
2. Moving from Workspace to Workspace
3. Executing Your Code Using the DBSTEP Function
4. Displaying Status Messages Periodically
5. Using the TRY/CATCH Block to Capture Errors
6. Using the ERROR Function with the LASTERR and RETHROW Functions
7. The WHICH Function

Topic 3 - Executing Your Code Using the DBSTEP Function
If you want to simply run your code, line by line, and verify the output after each step, you can use the DBSTEP function inside debug mode. Simply set a breakpoint (in the Editor/Debugger or use DBSTOP) at the line where you want to start your step-by-step investigation of the code and run the program.

For example, the file buggy.m, which is used to illustrate the debug functions, consists of three lines.

There are also three alternate syntaxes for DBSTEP, which will allow you to do more than simply step through one line:

  • dbstep N
    This will cause MATLAB to execute the next N lines, where N is a positive integer.
  • dbstep in
    If the next line to be executed is a call to another function, this syntax will cause the Editor/Debugger to open that function. MATLAB will stop at the first line in that function, as though there was a breakpoint at that line.
  • dbstep out
    This syntax will cause MATLAB to execute the rest of the lines in the current function, then return to the calling function and stop on the line immediately after the function call.

For more information on the DBSTEP functions, please visit the following link:

DBSTEP:
http://www.mathworks.com/access/helpdesk/help/techdoc/ref/index.html?/access/helpdesk/
help/techdoc/ref/dbstep.html


Checking your model with Simulink Model Advisor

The Model Advisor checks a model or subsystem for conditions and configuration settings that can result in inaccurate or inefficient simulation of the system represented by the model or generation of inefficient code from the model. It produces a report that lists all the suboptimal conditions or settings that it finds, suggesting better model configuration settings where appropriate.

Checking with the Model Advisor lets you avoid problems that arise from incomplete models or poor modeling style and reminds you of the recommended settings. You can customize the checks using Simulink Verification and Validation. For instance, you can check that:

  • All subsystem blocks have a blue background color
  • All Sum blocks are square
  • All library-linked blocks come from your company's custom libraries

Real-Time Workshop provides several Model Advisor demo models. To view them, type modeladvisor('rtwdemo_advisor1') at the MATLAB command line.

You can use an M-based API to customize the full contents of the Model Advisor. To see a customization demo, type slvnvdemo_mdladv at the MATLAB command prompt.

Using M scripting language and the Simulink API together with the expert advice of the Model Advisor, you can let your imagination and requirements guide you to more accurate, efficient simulations and well-constructed code.
 
 
EVENTS & TRAINING
 

Learn and do more with MATLAB & Simulink

'Building GUIs with MATLAB' Training Course

This 1-day hands-on course demonstrates how to build graphical user interfaces (GUIs) in MATLAB. It explains how to build GUIs in MATLAB from the command line by using MATLAB's graphical user interface development environment (GUIDE). The course also introduces concepts for designing and laying out GUIs. Demos show how actions defined by code can be linked to a user interface object, such as a push button. Explains how to create custom menus for GUIs

In-house or customized training is also available on request, please contact Activemedia at 6742-8173 for details. Other relevant training courses provided by Activemedia include:

- Comprehensive MATLAB
- Comprehensive SIMULINK
- Applying Signal Processing with MATLAB
- Applying Image Processing Techniques with MATLAB and SIMULINK
- Applying Neural Network with MATLAB
 
 
Visit www.activemedia.com.sg or Contact us at:
Singapore:
(65) 6742 8173
enquiry@activemedia.com.sg
Malaysia:
(60) 3 7880 8522
enquiry@activemedia.com.my
Thailand:
(66) 2 612 9390-1
info@activemedia.in.th
 
 
Customer Applications
 

Pacifica Group Technologies Advances Brake-by-Wire Technology with Simulink and xPC Target

To design an innovative brake-by-wire system without building and testing costly hardware prototypes
Use Simulink and xPC Target to design, prototype, and rigorously test a safety-critical brake-by-wire system for aggressive production schedules
• Huge savings in design and testing costs.
• Holistic testing.
• Faster prototyping and fewer field tests.
 

With brake-by-wire technology, we may soon be driving cars that decelerate automatically in hazardous situations, such as skidding on wet roads. Part of a larger trend in the automotive industry, brake-by-wire technology replaces mechanical linkages with electric signals, thereby reducing manufacturing costs and environmental concerns associated with hydraulic brakes. To date, however, implementing this technology has required building and testing costly and labor-intensive hardware prototypes.

Hardware-in-the-loop testing environment.
 

Pacifica Group Technologies (PGT), the R&D arm of Pacifica Group Limited, has bundled their resources into RABiT, a Victorian State Government-supported collaborative venture for advanced by-wire research. PGT used Simulink and xPC Target to design and test an innovative brake-by-wire system, greatly reducing their dependence on physical prototypes.

"With the progress we are making in this area, it is very likely we will be buying future cars with brake-by-wire technology features, and it wouldn't be happening without MathWorks tools," says Jaap Overschie, senior control systems design engineer at PGT."

Challenge

To preserve driver safety, brake-by-wire systems must be thoroughly reliable and fault-tolerant. PGT needed a way to test their technology under a broad range of driving conditions, and in the context of the entire vehicle system.

"Designing and testing brake-by-wire technology requires a scalable, integrated environment in which each part aligns with all other mechanisms of the vehicle," says Overschie. "Our challenge is to match the physical model of the new brake components with a real vehicle, and calibrate the model to identify any irregularities."

PGT faced intense global competition in developing these production-ready systems, so they needed a more cost-effective and faster way to test the system than traditional approaches. Typically, engineers developed a physical prototype and tested it in the field with real vehicles. After every test run, they adjusted or rebuilt the prototype, then tested it on the road again. This method was too costly and caused long time delays.

Moreover, purchasing and customizing a commercial off-the-shelf simulator with proprietary hardware and software would cost several hundred thousand dollars.

"Simulink and xPC Target give us an enormous cost savings. What we are designing and testing now would have cost 100 times more without MathWorks tools for Model-Based Design."

Jaap Overschie,
Pacifica Group Technologies
 

Solution

Instead of relying on conventional field testing approaches or expensive vehicle simulators, PGT designed, prototyped, and tested their brake-by-wire systems with MathWorks tools for Model-Based Design. They used Simulink for modeling, simulation, and analysis, and xPC Target for hardware-in-the-loop testing.

"We chose MathWorks products because they enable us to simulate and test every aspect of the brake system operation," says Overschie. "If we can simulate it, we can control it in real time.

PGT used Simulink to design a vehicle dynamics model that simulates vehicle response to the brake-by-wire products that they are developing. PGT used Real-Time Workshop to automatically generate C code that they ran on standard, low-cost personal computers with xPC Target. They used the PCs to simulate all vehicle components in real-time by running hardware-in-the-loop simulations with xPC Target.

"Using xPC Target is a very cost-effective way to test the internal software for model coverage, and external real-world influences through scenario testing," explains Overschie. "Conducting tests outside of such an environment is complicated, dangerous, and expensive."

If the tests run well, simulated components, such as a front right wheel, will be replaced with real components. They are integrated with the remaining simulated components through a real-time data link for hardware-in-the-loop testing. PGT can simulate the effect of a wheel hitting a curb, for instance, and the reaction of the real electromechanical brake.

PGT uses the Optimization Toolbox and the Genetic Algorithm and Direct Search Toolbox to automatically optimize and calibrate the design parameters of the new braking system. They evaluate these parameters on a real car, and then bring results back into MATLAB to further test and optimize a new set of parameters.

After rapid prototyping and verification, the team uses Real-Time Workshop Embedded Coder to transfer the desktop simulation into production C code for their ECU.

General Motors (GM) has incorporated PGT's brake-by-wire technology into the Sequel, a fuel cell car that debuted at the 2001 North American International Auto Show. GM says that the Sequel's state-of-the-art, by-wire systems will offer better control and more responsiveness than a conventional vehicle by slowing vehicles from 60 to 0 MPH in one car length, a stopping distance of about 5 meters shorter than conventional hydraulic brakes.
 

Results

  • Huge savings in design and testing costs. "Simulink and xPC Target give us an enormous cost savings," explains Overschie. "What we are designing and testing now would have cost 100 times more without MathWorks tools for Model-Based Design."
     
  • Holistic testing. "With Simulink and xPC Target, we can take a holistic approach to the car's operation, rather than viewing the components as separate units," Overschie says. "Now, we can truly assess the effect of our brake design changes on the entire vehicle."
     
  • Faster prototyping and fewer field tests. "MathWorks tools help us to prototype more rapidly," Overschie explains. "Field testing is now only required in the final stages of vehicle development."

Products Used

• Genetic Algorithm and Direct Search Toolbox
• MATLAB®
• Optimization Toolbox
• Real-Time Workshop®
• Real-Time Workshop® Embedded Coder
• Simulink®
• xPC Target