AI, deep learning, autonomous systems, and automotive toolboxes have been added in large numbers.


On the 22nd, MathWorks announced Release 2020b (R2020b) of MATLAB and Simulink.

MATLAB R2020b adds new bubble and swarm charts, the ability to compare and merge App Designer apps with the MATLAB Comparison Tool, and customizable figure icons and components.

Simulink R2020b introduces Simulink Online, which allows you to view, edit, and simulate Simulink models through a web browser, and the ability to generate code for model hierarchies referenced in Simulink up to 2x faster.

R2020b also includes new toolboxes built on AI capabilities, toolboxes to accelerate autonomous system development, and toolboxes to accelerate 3D scene creation for autonomous driving simulation.


Toolbox features added or updated in R2020b
The new 'Deep Learning HDL Toolbox' is aimed at algorithm developers and hardware designers.Provides functions and tools to help implement and prototype deep learning networks on A and SoC.

Additionally, it provides pre-built bitstreams that can run various deep learning networks on FPGA and SoC devices from Xilinx and Intel. It also provides profiling (performance measurement) tools that allow you to customize deep learning networks by exploring tradeoffs between design, performance, and resource utilization.

Using this toolbox, HDL Coder, and Simulink, engineers can customize the hardware implementation of deep learning networks and generate portable, synthesizable Verilog and VHDL code for deployment on FPGAs.

Starting with R2020b, Deep Learning Toolbox supports Image Classification and Network Prediction blocks to help generate code and simulate deep learning models in Simulink.

The 'Experiment Manager' app also enables deep learning network training in R parallel mode and hyperparameter tuning using Bayesian optimization.

The 'Deep Network Designer' app also adds the ability to train networks for image classification, semantic segmentation, multiple inputs, out-of-memory tasks, image-to-image regression, and other workflows.

The new Lidar Toolbox provides algorithms, functions, and apps for designing, analyzing, and testing Lidar processing systems, while the UAV Toolbox provides tools and reference applications for designing, simulating, testing, and deploying applications for unmanned aerial vehicles, or drones.

Updated 'Sensor Fusion and Trac Toolbox'The 'king Toolbox' includes the ability to track objects using a grid-based random finite set (RFS) tracker and to generate orbits using Earth-centered waypoints.

The new RoadRunner Scene Builder product, part of the RoadRunner product line specialized for the automotive industry, automates the creation of road networks from HD maps.

The AUTOSAR Blockset update now supports schema version 4.4 for importing and exporting ARXML files and generates AUTOSAR-compatible C code. It also provides a Linux executable for adaptive models, allowing you to create AUTOSAR adaptive executables and run them as standalone applications.

The Vehicle Dynamics Blockset now includes the ability to implement 6DOF trailers and vehicles with three axles, and Simulink 3D blocks to visualize tractors and trailers in the Unreal Engine 3D environment.

R2020b is available worldwide immediately. All new products, enhancements, and bug fixes in MATLAB and Simulink can be found on the R2020b Highlights page.