AI for Wireless

Apply artificial intelligence (AI) techniques to wireless communications applications

Whether you use machine learning, deep learning, or reinforcement learning workflows, you can reduce development time with ready-to-use algorithms and data generated with MATLAB® and wireless communications products. You can easily leverage existing deep learning networks outside MATLAB; streamline training, testing, and verification of your designs; and simplify deployment of your AI networks on embedded devices, enterprise systems, and the cloud.

With MATLAB, you can:

  • Generate training data in the form of synthetic and over-the-air signals using the Wireless Waveform Generator app
  • Augment signal space by adding RF impairments and channel models to your generated signals
  • Label signals collected from wireless systems using the Signal Labeler app
  • Apply reusable and streamlined training, simulation, and testing workflows to various wireless applications using the Deep Network Designer and Experiment Manager apps
  • Add custom layers to your deep learning designs

Why Use AI for Wireless?

Using a neural network to identify 5G NR and LTE signals in a wideband spectrogram.

Spectrum Sensing and Signal Classification

Identify signals in a wideband spectrum using deep learning techniques. Perform waveform modulation classification using deep learning networks.

Design a radio frequency (RF) fingerprinting convolutional neural network (CNN) with simulated data.

Device Identification

Develop radio frequency (RF) fingerprinting methods to identify various devices and detect device impersonators.

A screenshot of a spectrum analyzer shows that the performance characteristics change when the power amplifier (P A) heats, which creates a visual plot system as a function of time.

Digital Pre-Distortion

Apply neural network-based digital predistortion (DPD) to offset the effects of nonlinearities in a power amplifier (PA).

Comparing 5G NR channel estimates based on either idealized estimation, linear interpolation, or deep learning techniques.

Beam Management and Channel Estimation

Use a neural network to reduce the computational complexity in the 5G NR beam selection task. Train a CNN for 5G NR channel estimation.

Comparing actual locations of objects in a room with color-coded locations predicted using CNNs.

Localization and Positioning

Use generated IEEE® 802.11az™ data to train a CNN for localization and positioning.

Visualizing constellation plots of various autoencoders that converge to standard modulations such as Q P S K or 16 P S K.

Transceiver Design

Use an unsupervised neural network that learns how to efficiently compress and decompress data, forming an autoencoder. Train and test a neural network to estimate likelihood ratios (LLR).