Skip to Main Content Skip to Search
Accelerating the pace of engineering and science

 

Keynote Presentation

Technical Computing and Model-Based Design – 2011 and Beyond

Continuing advances in MATLAB and Simulink are creating new opportunities in technical computing and Model-Based Design across engineering and science. In this presentation, Jim discusses how technical computing with MATLAB takes advantage of ubiquitous data sources, more powerful computing hardware, and new analysis methods. He also describes recent advances in applying Model-Based Design for design exploration, code generation, and verification in the development of complex systems.

MATLAB and Simulink Updates

MATLAB Today

Discover how new MATLAB capabilities enable you to leverage multicore and GPU processors, design computer vision algorithms, and apply advanced statistical analysis techniques. The talk also covers new features in symbolic computing and data acquisition, tools for integrating MATLAB with Microsoft® Excel®, and generating C code automatically with the new MATLAB Coder.


Model-Based Design: From Concept to Production

Discover how new capabilities in the Simulink product family enable you to design, implement, and verify control, signal processing, and communications systems. Learn about new features for modeling multidomain systems, collaborating on large projects, and implementing embedded systems with automatic C, HDL, and PLC code generation. Find out about new products for model and code verification.

Technical Sessions

Applied Data-Fitting Techniques with MATLAB: Beyond the Basics

The MATLAB product family offers a variety of approaches for fitting and simplifying data. In this session, we demonstrate how you can:

  • Use linear and nonlinear regression to model data sets with outliers and clipping
  • Generate black-box models using boosted decision trees and localized regression
  • Reduce the dimensionality of data using feature selection, feature transformation, and shrinkage

After this session, you’ll understand the utility of the various techniques for parametric and non-parametric fitting, when to use each, and why.


Best Practices for Industry-Scale Modeling

This session highlights best practices for efficiently scaling up models. These best practices are based on MathWorks consultants’ and field engineers’ experience with helping customers adopt Model-Based Design within their development processes. This session is valuable to both customers working with large models and those who want to be able to support large models in the future.

MathWorks engineers demonstrate how to best architect and partition a design, define component interfaces, and work within configuration management systems. The session also provides resources you can use to help implement these best practices within your specific development processes.


Best Practices for Large-Scale Modeling

This session highlights best practices for efficiently scaling up models. These best practices are based on MathWorks consultants’ and field engineers’ experience with helping customers adopt Model-Based Design within their development processes. This session is valuable to both customers working with large models and those who want to be able to support large models in the future.

MathWorks engineers demonstrate how to best architect and partition a design, define component interfaces, and work within configuration management systems. The session also provides resources you can use to help implement these best practices within your specific development processes.


FPGA and ASIC Prototyping of Signal Processing Algorithms with MATLAB and Simulink

The “secret sauce” in many modern signal processing and communications systems consists of unique algorithms that are authored in MATLAB and Simulink. Engineers who are designing these algorithms are seeking out tools that facilitate rapid FPGA prototyping for early verification of their concepts. Since its introduction five years ago, many engineers have successfully used Simulink HDL Coder to create FPGA prototypes and tape-out ASICs.

In this session, we demonstrate how you can leverage automatic HDL code generation, HDL cosimulation, and FPGA-in-the-loop technologies to accelerate your FPGA prototyping workflow. We also illustrate how you can integrate these technologies into your existing FPGA and ASIC design workflows. You will learn how to:

  • Generate efficient, target-independent HDL code from Simulink models and MATLAB code
  • Verify HDL code through cosimulation with Mentor Graphics® ModelSim® and Cadence® Incisive®
  • Perform FPGA-in-the-loop verification with Simulink

Mathematical Modeling with MATLAB

The MATLAB product family enables you to build mathematical models for forecasting and optimizing the behavior of complex systems. In this session, we demonstrate how you can:

  • Develop models using data fitting and first-principles modeling techniques
  • Simulate models and develop custom postprocessing routines
  • Generate reports that document models and simulation results

The session covers use of the MATLAB language, symbolic expressions, and prebuilt graphical tools for specific modeling tasks and other approaches you can use to develop models.


Model-Based Design for DO-178

This session focuses on Simulink support for the DO-178B standard. It describes development and verification workflows in the context of objectives specified in DO-178. It presents new tools and capabilities that automate workflow activities, including source code verification. The session shows how to use DO Qualification Kit to qualify tools and concludes with a brief discussion of Model-Based Design for DO-178C.


Model-Based Design for High-Integrity Applications

This session focuses on Simulink support for high-integrity applications. It describes development and verification workflows necessary for developing robust and verifiable software using Model-Based Design. It presents new tools and capabilities that automate workflow activities, including source code verification. The session concludes by introducing certification and qualification kits that facilitate use of MathWorks tools for DO-178, IEC 61508, ISO 26262, and other popular safety standards.


Parallel Computing with MATLAB

This session focuses on how you can improve the performance of your MATLAB code using MathWorks parallel computing products. You will learn how to speed up your applications with minimal programming effort using widely available multicore desktop systems. We highlight parallel programming constructs such as parallel for-loops, as well as the management of jobs and tasks.

Master Classes

Computer Vision with MATLAB

In this master class, we discuss the new Computer Vision System Toolbox as well as Image Acquisition Toolbox, Image Processing Toolbox, and Statistics Toolbox. Through product demonstrations, you will learn how to use capabilities for feature detection, extraction, matching, and classification to solve computer vision challenges. We also demonstrate how to read, write, process, and display video.


Design, Generate, and Verify C Code for Your Embedded Controller Using Simulink

This master class shows the basic workflow for designing, generating, and verifying C code for a control algorithm. The class features a field-oriented controller for an AC motor using a TMS320F28335 MCU from Texas Instruments to demonstrate this workflow; the tools and techniques are easily extended to other applications and devices.

The class is structured in three sections:

  • Design and test functional requirements through simulation
  • Generate controller code and integrate with application software
  • Verify functional and resource requirements of compiled code

Functional verification of compiled code is accomplished using processor-in-the-loop (PIL) testing. You will see how PIL testing enables you to verify that the compiled executable matches the model behavior. You will also see how PIL testing enables you to profile code execution time to assess resource utilization. An example of integrating a processor optimized library is used to demonstrate verification of functional and resource requirements for a design change.


Generate, Verify, and Integrate C Code from Your MATLAB Algorithms

This master class walks you through the capabilities of MATLAB Coder and DSP System Toolbox for generating, verifying, and integrating readable and portable C code from your MATLAB algorithms for signal processing applications. You’ll see examples that show how to introduce implementation constraints to your MATLAB algorithms to prepare them for code generation, and then generate MEX-files for verifying the compiled behavior and accelerating algorithm execution. Finally, you’ll see how to generate readable and customizable C code and integrate it with other software such as a Microsoft® Visual Studio® parent project.


Parallel Computing

In this master class, we demonstrate how to boost execution speed on computationally intensive problems in MATLAB. We begin by highlighting add-on tools you can use to take advantage of parallel computing resources without requiring any extra coding. We then introduce parallel programming constructs and address how to overcome the memory limits of your desktop computer and work with very large arrays. You will learn how to take advantage of common computing trends including GPUs, multiprocessor machines, computer clusters, grids, and clouds. You will discover how Simulink can run in parallel. Lastly, we show how you can deploy your parallel applications to other users.

This session covers:

  • Using toolboxes with built-in support for parallel computing
  • Speeding up computations using additional cores or GPUs
  • Working with larger data sets
  • Scaling up to computer clusters, grids, and clouds
  • Developing and deploying your parallel algorithms