what’s new in matlab and simulink - se.mathworks.com file– visualize and denoise time series...
TRANSCRIPT
1© 2015 The MathWorks, Inc.
What’s new in MATLAB and
Simulink
Mandar Gujrathi
2
ApplicationBreadth
Products for the work you do
WorkflowDepth
Support for your entire workflow
PlatformProductivity
Getting your work done faster
3
ApplicationBreadth
Products for the work you do
WorkflowDepth
Support for your entire workflow
PlatformProductivity
Getting your work done faster
WorkflowDepth
PlatformProductivity
ApplicationBreadth
▪ Simulate Faster
▪ Create Your Designs Faster
▪ Simplify Analysis
▪ Scale Your Work
▪ Collaborate
4
Simulate Faster
▪ All MATLAB code can now be JIT compiled
▪ MATLAB runs your code over twice
as fast as it did just three years ago
▪ No need to change a single line of your code
▪ Increased speed of MATLAB startup in R2018a
Redesigned execution engine
runs MATLAB code faster
5
Create Your Designs Faster
MATLAB
Live Editor
6
Create Your Designs Faster
testCase.press(myApp.checkbox)
testCase.choose(myApp.discreteKnob, "Medium")
testCase.drag(myApp.continuousKnob, 10, 90)
testCase.type(myApp.editfield, myTextVar)
MATLAB
App Designer
7
Create Your Designs Faster
MATLAB Simulink
8
Create Your Designs Faster
MATLAB Simulink Stateflow
9
Simplify Analysis
▪ Econometric Modeler app
– Perform time series analysis, specification
testing, modeling, and diagnostics
▪ Analog Input Recorder app
– Acquire and visualize analog input signals
▪ Wavelet Signal Denoiser app
– Visualize and denoise time series data
These interactive applications automate
common technical computing tasks
Econometrics ToolboxData Acquisition ToolboxWavelet Toolbox
Simplify Analysis with Apps
11
Scale Your Work
▪ Run multiple parallel simulations with parsim
▪ Monitor simulation status and progress in the
Simulation Manager
Use parallel computing to run multiple
simulations faster
Parallel Computing ToolboxMATLAB Distributed Computing Server
12
Scale Your Work
▪ Use familiar MATLAB functions and syntax
▪ Support for hundreds of functions
▪ Works with Spark + Hadoop clusters
Use tall arrays to manipulate and analyze
data that is too big to fit in memory
Statistics and Machine Learning Toolbox
14
Team Collaboration
Use advanced software development features
to manage, test, and integrate MATLAB code
Identify differences between model
elements, Stateflow charts, and
MATLAB Function blocks
15
WorkflowDepth
PlatformProductivity
ApplicationBreadth
▪ Create Your Designs Faster
▪ Simplify Analysis
▪ Simulate Faster and Scale Your Work
▪ Collaborate
16
WorkflowDepth
PlatformProductivity
ApplicationBreadth
▪ Deployment of MATLABAlgorithms and Applications
▪ Code Generation fromSimulink Models
▪ Verification and Validation
17
Deploy MATLAB Algorithms and Applications
Analyze Data
Data
exploration
Preprocessing
Domain-specific
algorithms
Access Data
Sensors
Files
Databases
Deploy
Embedded
devices
Desktop apps
Enterprise
systems
Develop
AI model
Algorithm
development
Modeling &
simulation
18
MATLAB CompilerMATLAB Compiler SDKMATLAB Production Server
Deploy MATLAB Algorithms and Applications
▪ Standalone desktop applications
▪ Add-ins for Microsoft Excel
▪ Software components to integrate with other languages (C/C++, .NET, Python, Java)
▪ Software components for web and
enterprise applications
Share your work outside of MATLAB without
having to recode your algorithms
Analyze Data
Data exploration
Preprocessing
Domain-specific algorithms
Access Data
Sensors
Files
Databases
Deploy
Embedded devices
Desktop apps
Enterprise systems
Develop
AI model
Algorithm development
Modeling & simulation
19
MATLAB CompilerMATLAB Compiler SDKMATLAB Production Server
Deploy MATLAB Algorithms and Applications
▪ Standalone desktop applications
▪ Add-ins for Microsoft Excel
▪ Software components to integratewith other languages
(C/C++, .NET, Python, Java)
▪ Software components for web and
enterprise applications
Share your work outside of MATLAB without
having to recode your algorithms
Analyze Data
Data exploration
Preprocessing
Domain-specific algorithms
Access Data
Sensors
Files
Databases
Deploy
Embedded devices
Desktop apps
Enterprise systems
Develop
AI model
Algorithm development
Modeling & simulation
MATLAB Production Server
Request
BrokerMobile / Web
Application
< >
3rd Party
Dashboard
Enterprise
Application
Data Sources
Edge Devices
Analytics Development
20
Deploy MATLAB Algorithms
▪ Generate C code for predictive machine learning
and deep learning models
▪ Generate optimized CUDA code for deep learning,
embedded vision, and autonomous systems
Deploy machine learning and deep learning
models using automatically generated code
MATLAB CoderGPU Coder
Analyze Data
Data exploration
Preprocessing
Domain-specific algorithms
Access Data
Sensors
Files
Databases
Deploy
Embedded devices
Desktop apps
Enterprise systems
Develop
AI model
Algorithm development
Modeling & simulation
21
PID Control Tuning
▪ Automatically tune PID controller gains in
real time against a physical plant
▪ No model of plant dynamics required
▪ Deploy the auto-tuning algorithm to
embedded software using automatic code
generation
Implement an embedded PID
auto-tuning algorithm
Simulink Control Design
25
Connecting Your Design to Hardware
▪ Live streaming to and from hardware
▪ Run Simulink models on low-cost hardware,
such as Arduino, Raspberry Pi, and LEGO
▪ Automatically generate code and run it on
microprocessors, FPGAs, and more.
Connect directly to hardware with
support packages
Arduino
Raspberry Pi Microsemi FPGA
ADALM-PLUTO
ARM Cortex
LEGO
26
Manual architecture
design & coding
Specification doc
Deploying to FPGA or ASIC Hardware
HDL VerifierHDL CoderFixed-Point DesignerVision HDL ToolboxLTE HDL Toolbox
Algorithm
Fixed-Point HDL
FPGA/ASIC
Implementation
Algorithm w/ Hardware
Implementation
HDL
CoderHD
L V
eri
fier
Vision HDL Toolbox LTE HDL Toolbox
Matrix SupportNative Floating Point
HDL Checks in Model Advisor
28
WorkflowDepth
PlatformProductivity
ApplicationBreadth
▪ Deployment of MATLABAlgorithms and Applications
▪ Code Generation fromSimulink Models
▪ Verification and Validation
29
WorkflowDepth
PlatformProductivity
ApplicationBreadth
▪ Autonomous Systems
▪ Artificial Intelligence (AI)
▪ Wireless Communications
30
Designing Autonomous Systems
Sense
PerceiveDecide
& Plan
Act
31
Sense
PerceiveDecide
& Plan
Act
Sense
PerceiveDecide & Plan
Act
Computer Vision System ToolboxRobotics System Toolbox
▪ Segment and register lidar point clouds
▪ Lidar-Based SLAM: Localize robots and
build map environments using lidar sensors
Mapping of environments using sensor data
Designing Autonomous Systems
32
Designing Autonomous Systems
▪ Object detection and tracking
▪ Semantic segmentation using deep learning
Understanding the environment using computer
vision and deep learning techniques
Sense
PerceiveDecide & Plan
Act
Neural Network ToolboxComputer Vision System ToolboxAutomated Driving System Toolbox
CamVid Database: Brostow, Gabriel J., Julien Fauqueur, and Roberto Cipolla. "Semantic object classes in video: A high-definition ground truth database." Pattern Recognition LettersVol 30, Issue 2, 2009, pp 88-97.
33
Designing Autonomous Systems
▪ Interactively design synthetic driving scenarios
composed of roads and actors
(vehicles, pedestrians, etc.)
▪ Generate visual and radar detections of actors
Design synthetic driving scenarios to test
controllers and sensor fusion algorithms
Sense
PerceiveDecide & Plan
Act
Automated Driving System Toolbox
Driving Scenario Designer App
35
Full Vehicle Simulation
Vehicle Dynamics Blockset
Ride & handling Chassis controls Automated Driving
Sense
PerceiveDecide & Plan
Act
36
Artificial Intelligence
COMPUTER
Output
Data
Model
37
Text Analytics
Text Analytics Toolbox
Output
Data Model
38
Text Analytics
▪ Preprocess raw text data by
extracting, filtering, and splitting
▪ Visualize text using word clouds
and text scatter plots
▪ Develop predictive models using
built-in machine learning algorithms
(LDA, LSA, word2vec)
Work with text from equipment logs
and operator reports
Text Analytics Toolbox
39
Normal Operation Maintenance NeededMonitor Closely
Predictive Maintenance
Predictive Maintenance Toolbox
Sensors
Output
Data Model
40
Predictive Maintenance
Predictive Maintenance Toolbox
▪ Import sensor data from local files and cloud storage(Amazon S3, Windows Azure Blob Storage, and Hadoop HDFS)
▪ Use simulated failure data from Simulink models
▪ Estimate remaining useful life (RUL)
▪ Get started with examples
(motors, gearboxes, batteries, and other machines)
Design and test condition monitoring and
predictive maintenance algorithms
41
Deep Learning
Neural Network ToolboxComputer Vision System ToolboxGPU Coder
Output
Data Model
42
Deep Learning
▪ Access the latest models
▪ Import pretrained models and
use transfer learning
▪ Automate ground-truth labeling using apps
▪ Design and build your own models
▪ Use NVIDIA GPUs to train your models
▪ Automatically generate high-performance
CUDA code for embedded deployment
Design, build, and visualize convolutional neural networks
Neural Network ToolboxComputer Vision System ToolboxGPU Coder
AlexNet ResNet-50 VGG-16
TensorFlow
MATLAB
MXNet
GPU Coder
Imag
es
/ se
c
43
44
Design with the Latest Wireless Standards
802.11ax
NB-IoT
46
RF and Antenna Design and Prototyping
▪ RF top-down design with RF Budget Analyzer app
▪ Adaptive hybrid beamforming and MIMO system modeling
▪ RF Power Amplifier modeling and DPD linearization
▪ RF propagation and 3D terrain visualization
▪ Design and fabrication of printed (PCB) antennas
Use RF and Antenna models through
your entire development cycleFrom idea …
… to implementation
????
?
???
?? ?
??? ?
Antenna ToolboxRF ToolboxRF Blockset
48
▪ Deployment
▪ Code Generation
▪ Verification and Validation
WorkflowDepth
What’s New in MATLAB and Simulink?
▪ Design Creation
▪ Analysis
▪ Simulation, Scaling
▪ Collaboration
PlatformProductivity
▪ Autonomous Systems
▪ Artificial Intelligence (AI)
▪ Wireless Communications
ApplicationBreadth
49
Upgrade your MATLAB Code and Simulink Models
50© 2015 The MathWorks, Inc.