Military Embedded Systems

Handling the information overload – and the heat

Story

August 04, 2020

John McHale

Editorial Director

Military Embedded Systems

Information overload: another way of saying big data challenge.

In other words, there is way too much information out there for military operators to sift through – whether it’s sifting through signals intelligence, video feeds from unmanned aircraft, or text such as social media posts from suspected terror groups. The amount of data gathered is so massive that it’s impossible to properly search it all in a reasonable amount of time.

Even more daunting: No matter how large that pile of data might be, we still don’t have enough data gathered.

Depressing, for sure, but not hopeless. Technological solutions are being leveraged to gather that data. We see it in our everyday lives, for example when shopping online. The other day I looked at a new golf club on a manufacturer’s website and the ad for that club followed me to Facebook, to LinkedIn, all over the place. Ever wonder how Barnes and Noble knows what you like to read? The company mines the data behind your book purchases and makes recommendations to you for your next book based on your data.

But that way of using data is kind of old hat. More urgently, big data also figures into the challenge the medical world and the government have in tracking the spread of the COVID-19 virus. That set of critical tasks may actually more closely mirror the challenge faced by military regarding big data. Both are matters of life and death – both require new information every day – both will never ever truly have enough data to fight the virus or fight the enemy without casualties.

While it is true that commercial applications and game systems have acknowledged this challenge and apply big data in creative ways, “these industries do not operate in the life-and-death environments that warfighters deal with daily,” says Dr. Scott Neal Reilly, senior vice president and principal scientist at Charles River Analytics in a feature on page 32 by associate editor Emma Helfrich. “Developing data-driven systems that can be validated to the extent necessary for deployment in real-world, military contexts is a major challenge for the community.”

One major tool the U.S. Department of Defense (DoD) is using to solve the big data problem is artificial intelligence (AI), which Helfrich also covers in her feature.

“The DoD AI strategy is prioritizing systems that reduce cognitive overload and improve decision-making,” says Michael Rudolph, aerospace and defense industry manager for MathWorks, in the feature. “In order to speed up that OODA [observe, orient, decide, act] loop, AI needs to make earlier predictions and identify emerging issues from a variety of data sources. With big data in AI, the question is not just how much data but one of data and feature quality.”

Much of the big data challenge is being undertaken by software engineers, those specializing in AI and machine learning. However, performing AI functions in software requires very powerful computers and processors, which brings us to a problem that military embedded system designers have been facing for decades – thermal management.

Every year, processor performance gets faster and more impressive, but the heat they generate remains a problem military commercial off-the-shelf (COTS) suppliers must solve through such techniques as air-flow-through cooling, liquid cooling, or conduction cooling. Solving that speed/heat problem never gets easier, especially in VPX-based systems.

“With the advent of individual 6U VPX modules that surpass 200 watts of dissipated power, traditional conduction cooling approaches for heat removal are being pushed to the limit and with that we see the emergence of the various VITA 48 module level cooling strategies,” say Steve Gudknecht and Jordan Sudlow of LCR Embedded Systems in their article on page 22. The authors go into an overview of different cooling techniques in the piece.

In our Industry Spotlight on thermal management, Max Taylor-Smith of Entropy Electro-Mechanical Solutions takes a very in-depth approach on how to simulate the thermal challenges inherent with FPGAs. While these devices enable unprecedented performance for signal processing applications, they are also notorious for turning up the heat as they blaze through reams of data.

In his piece, Taylor-Smith observes that the approach to thermal solutions for higher-power FPGA versions has remained relatively stagnant for the last 10 years. Read his take on this situation on page 26.

Another not-to-miss article on beating the heat: Pat Quinn of Curtiss-Wright Defense Solutions illustrates the ways in which designers can manage thermal issues that arise in data-acquisition units for extreme environments, like those used for flight test instrumentation. Read this one on page 18.

Categories
A.I. - Big Data
Topic Tags