MathWorks releases 2019b MATLAB and Simulink to help advance AI and automotive

MathWorks announced the release of Release 2019b, which includes a range of new features for MATLAB and Simulink, including support for artificial intelligence, deep learning, and the automotive industry. In addition, R2019b introduces new products that support robotics, new training resources for event-based modeling, and updates and bug fixes for the MATLAB and Simulink product lines. Release highlights include:

MATLAB

MATLAB highlights in R2019b include the introduction of Live Editor tasks, which allow users to interactively explore parameters, preprocess data, and generate MATLAB code as part of a Live Script. Now, MATLAB users can focus on the task itself, not syntax or complex code, and can automatically run the generated code to quickly iterate on parameters through visualization.

Simulink

Simulink highlights in R2019b include the new Simulink Toolstrip, which helps users access and discover the functionality they need. In the Simulink Toolstrip, tabs are arranged according to workflow and sorted by frequency of use, saving navigation and searching time.   

Artificial Intelligence and Deep Learning

In R2019b, Deep Learning Toolbox builds on the flexible training loops and networks introduced earlier this year. New features enable users to train advanced network architectures with custom training loops, automatic differentiation, shared weights, and custom loss functions. In addition, users can now build generative adversarial networks (GANs), Siamese networks, variational autoencoders, and attention networks. Deep Learning Toolbox can also now export to ONNX format networks that combine CNN and LSTM layers, as well as networks that include 3D CNN layers.

car

R2019b also introduces significant support for the automotive industry across multiple products, including:

Automated Driving Toolbox: 3D simulation support, including the ability to develop, test, and validate driving algorithms in a 3D environment; and a module that enables users to generate velocity profiles of driving paths given kinematic constraints.

Powertrain Blockset: Generates deep learning SI engine models for algorithm design and performance, fuel economy, and emissions analysis. New fully assembled models, including the HEV P0, P1, P3, and P4 reference applications, are available for HIL testing, trade-off analysis, and control parameter optimization of hybrid electric vehicles. 

Sensor Fusion and Tracking Toolbox: Enables you to perform track-track fusion and build decentralized tracking systems. 

Polyspace Bug Finder: Expands support for AUTOSAR C++14 coding guidelines, checking for misused lambda expressions, potential enumeration issues, and other issues.

robot

In addition to new features in Robotics System Toolbox, R2019b introduces two new products:

Navigation Toolbox (new): allows you to design, simulate, and deploy algorithms for planning and navigation. It includes algorithms and tools for designing and simulating systems that can map, localize, plan, and move in physical or virtual environments. 

ROS Toolbox (new): can be used to design, simulate, and deploy ROS-based applications. The toolbox provides an interface between MATLAB and Simulink and the Robot Operating System (ROS and ROS2), allowing users to build a node network, model and simulate the ROS network, and generate embedded system software for ROS nodes.  

Stateflow Training

The Stateflow Getting Started interactive tutorial is available in R2019b to help users learn the basics of how to create, edit, and simulate Stateflow models. Like the existing MATLAB, Simulink, and Deep Learning Getting Started tours, this self-paced learning course includes video tutorials and hands-on exercises with automatic assessments and feedback.

R2019b is available now worldwide.