AI Moves into Military Boards and Subsystems
Artificial intelligence (AI) is inching into high-end military markets, according to the suppliers of computer and sensor boards. However, challenges like meshing AI’s rapid advances with lengthy military contract times may slow widespread adoption. Developers of boards and subsystems at this year’s Embedded Tech Trends conference in Atlanta detailed the growing role of AI as well as challenges that must be overcome before it can achieve widespread usage. The conference is sponsored by VITA, a manufacturers group that creates many of the computer board standards used in military equipment.
“AI is redefining global power,” said Devon Yablonski, sensor products manager at Mercury Systems. “Over the last year, the government has invested a lot in this industry, trying to help us bring technologies from consumer into military applications. We’re seeing a lot of interest in the use of AI.”
Increased computing power is a key factor driving acceptance in all markets. Processing data for use with any of the many types of AI requires a lot of data handling and data manipulation. Currently, graphic processing units and FPGAs (field-programmable gate arrays) are used for the parallel tasks of these programs, which are already widely used in many fields. Those processors are also well suited to image processing, which is a key application for AI analysis.
“AI is good for anything with imaging, where you need to immediately determine what the image is," said Nigel Forrester, director of business development at Concurrent Technologies. “We’re supporting a number of neural networks. If CPU modules don’t provide enough processing power, you can add FPGA boards.”
Semiconductor technologies are advancing quickly, but neural networks and AI are evolving even more rapidly. These technologies are being used in areas like consumer products and data centers, providing huge markets that attract investment funding to fuel development teams. Defense contractors are leveraging the rapid advances.
“It is creeping into everything, it’s going to define most of the next big trends,” said John Bratton, sensor product marketing director at Mercury Systems. “It’s outpacing Moore’s Law, capabilities double every three to six months. Nobody patents anything because it’s obsolete in less than a year.”
That quick rate of change may pose problems for military programs that move on far longer development cycles. Many of the high-end systems and components developed by VITA members have long development and production periods. “Success comes over a long cycle,” said Shan Morgan, president of Elma Electronic. “It takes about 18 months to deliver a prototype, and it can be three to four years before you go to production.”
That’s not the only obstacle for broader usage of AI. While usage in leading-edge military systems is growing, more mainstream applications are emerging at a slower pace. Some of the challenges include verification, which can be difficult since system responses can change as systems learn. Another is the sea change for programmers who must learn complex new technologies that are rapidly changing.
“A lot of people don’t know how to use AI,” said Noah Donaldson, CTO at Annapolis Micro Systems. “They’re used to telling the computer what to do, not letting the computer do something programmers may not have thought of.”
While shipments may be relatively small now, most conference presenters are looking into AI’s potential roles in their strategies. Research programs are under way at companies as diverse as Interface Concepts, a French board manufacturer, and Elma, which makes the card cage packages that house computer boards. Some environments are well suited to AI’s analytical capabilities.
“Today, we don’t see a lot of people using AI,” said Michael Slonosky, at Curtiss-Wright. “It’s very useful for prognostics, predicting when failures might occur. Machine vision is also a good application for AI.”
The technology can bring significant benefits for warfighters. “When you don’t have a GPS connection but you do have stored maps and cameras, AI can help identify the area you’re in, in daylight or at night,” said David Saar, CTO at Aitech Defense Systems.
University of Rochester Lab Creates New 'Reddmatter' Superconductivity Material...
INSIDERElectronics & Computers
MIT Report Finds US Lead in Advanced Computing is Almost Gone - Mobility...
Airbus Starts Testing Autonomous Landing, Taxi Assistance on A350 DragonFly...
Boeing to Develop Two New E-7 Variants for US Air Force - Mobility Engineering...
PAC-3 Missile Successfully Intercepts Cruise Missile Target - Mobility...
Air Force Pioneers the Future of Synthetic Jet Fuel - Mobility Engineering...
Driver-Monitoring: A New Era for Advancements in Sensor Technology
Manufacturing & Prototyping
Tailoring Additive Manufacturing to Your Needs: Strategies for...
How to Achieve Seamless Deployment of Level 3 Virtual ECUs for...
Specifying Laser Modules for Optimized System Performance
Electronics & Computers
Leveraging Machine Learning in CAE to Reduce Prototype Simulation and Testing
University of Rochester Lab Creates New 'Reddmatter' Superconductivity Material
INSIDERTest & Measurement
New Consortium to Develop Thermal Protection Materials for Hypersonic Vehicles
Multi-Agent RF Propagation Simulator
Low Distortion Titanium in Laser Powder Bed Fusion Systems
How to Test a Cognitive EW System