WHITE PAPER
Processing Evolution for the Future Electronic Battlespace
mrcy.com 3
AI-ENHANCED RF APPLICATIONS:
CAPABILITIES AND CHALLENGES
New AI-based technologies offer tremendous promise
of new capabilities and advantages for the electronic
battlefield. Cognitive Radar applies AI techniques to
extract information about a target from a received
signal and then uses that information to improve
transmit parameters such as frequency, waveform,
and pulse repetition frequency. The feedback between
receiver and transmit functions differentiates a
Cognitive Radar from a standard radar, improving its
effectiveness. However, to be of practical use, this
feedback must be delivered with extremely low latency.
In a similar fashion, Cognitive EW applies AI to identify
patterns in detected RF signals and then develop an
appropriate response. The goal is to detect and capture
signals from adaptive, agile, stealthy, and out-of-library
transmitters, which is very difficult in congested RF
environments. As with Cognitive Radar, this AI-supported
pattern identification needs to happen in real time.
Executing AI algorithms in real time on a SWaP-
constrained platform presents a significant technical
challenge. Until recently, application SWaP constraints
made deploying AI-based systems untenable. There
are now two dominating approaches to support
cognitive radar/EW and AI inference at the edge:
Versal® ACAP with AI Core signal processing and GPU
vector processing. Versal ACAP (advanced compute
acceleration platform) is a software-programmable,
heterogeneous compute platform that combines
multiple styles of processing; while GPU computing
engines are excellent at executing the massively
parallel vector math operations that underlie AI. To
meet the myriad of AI application requirements across
the electronic battlefield, Mercury supports both
processing technologies as well as SiP implementations
for maximum performance and greater efficiency
THE BENEFITS OF DIRECT DIGITIZATION
Direct digitization avoids down converting a
signal to a lower intermediate frequency (IF)
before digitization; the ADC (analog to digital
converter) digitizes RF signal directly and sends
it on to the next step in the processing chain.
The primary benefit is a less complex RF signal chain,
with fewer analog components. This reduces the
number of potential noise sources and simplifies signal
synchronization, while also reducing per channel
costs. But for RF processing deployments, the most
significant result is in improved SWaP characteristics.
Nowhere is this more important than in an active
electronically scanned array (AESA), a type of phased
array radars. These systems form beams by phase
shifting the signals emitted from up to hundreds or even
thousands of antennas elements, with each element
supported by a transceiver. Keeping the transceiver
components as small and low-powered as possible is
essential for deployment, especially in airborne platforms.
Direct digitization delivers huge benefits to these designs.
Figure 2: SWaP-constrained airborne AESA radar system
