DEEPRadio – Deep Learning Over the Air

DeepRadio is a stand-alone radio platform providing the necessary speed and fidelity for intelligent decision making in spectrum learning and communications. DeepRadio deploys deep neural networks on the FPGA for low-latency and low-power characterization of the RF spectrum environment including 5G and adaptation to spectrum dynamics, channel interference, and congestion effects in real time (microseconds).

Smart Radio Empowered by Deep Learning

DeepRadio is a stand-alone radio platform providing the necessary speed and fidelity for intelligent decision making in spectrum learning and communications. DeepRadio deploys deep neural networks on the FPGA for low-latency and low-power characterization of the RF spectrum environment including 5G and adaptation to spectrum dynamics, channel interference, and congestion effects in real time (microseconds).

Low Latency, and Power Efficiency
DeepRadio is built upon FPGA empowered reconfigurable implementation of deep neural networks, offering low latency and low power consumption for a wide range of spectrum applications.

Truly Cognitive Communications
DeepRadio provides configurable deployment of full network protocol stack for effective interference management, spectrum adaptation, and high data rate communications.

RF Signal Characterization
DeepRadio uses deep learning to detect RF signals,
including 5G, in real time with embedded computing.

Downloads

Features

Low-Latency

Operates with low-latency (in microseconds) matching the speed of dynamic (interference, channel and traffic) changes in the spectrum.

Low-Power

Delivers high performance with low-power operation resulting in a longer battery lifetime (translated to longer network lifetime).

High Fidelity

Provides high fidelity by training on an extensive spectrum database with a diverse set of waveforms and channel conditions (path loss, fading, and multipath).

MIMO Communications

Offers adaptive modulation, space-time block coding, precoding, and beamforming for MIMO communications.

Field Tested

Field tested and validated for improving spectrum situational awareness and network performance.

Use Cases

Spectrum Agility

Classifies RF signals (received by RF front ends) quickly and accurately, providing a better characterization of the RF spectrum, including 5G, in real-time, supporting dynamic spectrum access (DSA), beamforming and MIMO systems.

Spectrum Co-Existence

Supports effective interference management by identifying types of signals and radio protocols, and sharing the spectrum while sustaining in communication quality.

Spectrum Management

Supports spectrum management by identifying interference to spectrum resources such as cellular base stations and WiFi access points, thereby allowing more robust planning of spectrum coverage and optimizing spectrum usage.

Perimeter Security

Identifies RF intrusions by recognizing unpermitted signals for surveillance applications.