May 17, 2019 by Sam Holland
Late March, MathWorks released the latest version of Simulink and MATLAB software: version R2019a. In this article, we look at the newest features available in the new update.
Simulink is an interactive programming environment used to model, simulate, and verify multi-domain dynamical systems. It is widely used in engineering and science communities across the globe for control automation, signal processing, machine learning, and more.
Check out some of the more prominent enhancements to the MATLAB and Simulink environments below.
Simulink R2019a allows engineers to expand their AI skill set with its new Reinforcement Learning Toolbox which offers interoperability with other deep learning frameworks such as Open Neural Network Exchange (ONNX), TensorFlow, and Keras.
According to MATLAB’s marketing director, David Rich, AI applications for the update include controllers, decision-making systems, and deep learning models on Nvidia DGX. Engineers can use the Deep Network Designer app to build, edit, analyse, and visualise deep artificial neural networks and generate the corresponding C/C++ codes to model and simulate AI systems. Also, users are granted more functionality than ever before—with computer vision, data acquisition, and image acquisition toolboxes.
The Computer Vision toolbox allows design and testing of video processing, computer vision, and 3D vision systems using a variety of operations such as ground truth labelling, camera calibration, 3D reconstruction, lidar/3D cloud processing and many more.
The Data Acquisition Toolbox can be used to configure data acquisition hardware by porting the data into MATLAB and Simulink and writing it to DAQ output hardware, such as USB, PCI/PCI Express, and PXI/PXI Express, while the Image Acquisition Toolbox handles image and video fetching and connects cameras and lidar sensors to MATLAB and Simulink.
Polyspace Bug Finder tool identifying source code defects. Image courtesy of MathWorks.
Version R2019a improves on polyspace static analysis for MATLAB and Simulink, including new products that enable enterprise-level use for safety and business-critical applications.
Products which implement polyspace static codes can be verified for model and performance integrity by using formal techniques for C/C++ and Ada to prove that there are no critical runtime errors under all possible control and data workflows. Three polyspace analysis products are included in the update: Polyspace Prover, Polyspace Bug Finder, and Polyspace for Ada.
The Polyspace prover formally verifies that there are no critical runtime errors in the C/C++ source codes without executing the program, and it can be used alongside Eclipse IDE on desktop. Polyspace Bug Finder scans a software’s source code to identify runtime errors, improper coding rules, and security vulnerabilities and fixes bugs wherever they are found.
Polyspace Client for Ada and Polyspace Server for Ada use static code analysis to formally prove the absence of critical runtime errors in Ada source code. All polyspace static analysis features are complete with quality software metrics tools for performance monitoring and compliance with industry standards.
Test bench for an analogue-to-digital converter. Image courtesy of MathWorks.
R2019a offers several enhancements for wireless and electronic signal processing via add-ons such as the Mixed Signal Blockset, SoC Blockset, and SerDes Toolbox.
The Mixed-Signal Blockset is used for rapid model construction and simulation of analogue and mixed-signal systems. It allows engineers to use models of components and impairments, test benches, DSP algorithms, and control logic to design and simulate phase-locked loops (PLL), analogue-to-digital converters (ADCs), and other systems at various levels of abstraction.
The SoC Blockset can be used to model, simulate, and implement hardware and software architecture for system-on-chip, FPGA, and ASIC with Simulink blocks and visualisation tools.
R2019a’s SerDes (Serializer/Deserializer) Toolbox provides a SerDes Designer app for rapid design, modelling, and analysis of wired communications transmitters and receivers. All data processing features are complete with data analysis and visualisation tools for error-detection, debugging, and performance monitoring.
An additional feature is the Aerospace Blockset, which allows aerospace engineers to model, visualise, and simulate aerospace vehicle dynamics, aiding rapid flight prototyping and reducing the design-to-testing period.
Engineers working in diverse industries rely on MATLAB and Simulink tools for product modelling and simulation. Version R2019a improves on its artificial intelligence, polyspace static analysis, and signal processing functions. The latest updates implement several bug fixes for improved stability and a range of features and enhancements engineers can utilise in their projects.
