The U.S. Air Force and every other branch of U.S. military service depend on safe and reliable microelectronic devices and embedded systems. Without them, modern defense weapons systems can’t function. And when the integrity of these components is compromised, it puts people, systems and entire military operations in peril.
The microelectronics industry is a broad and complex ecosystem of global supply chains that involves many different players, places and steps — and that won’t be changing anytime soon. In the past, our government led technological advancements in microelectronics, but various policies and regulations have driven up the cost of manufacturing devices in the U.S. and pushed that part of the process overseas. Subsequently, microelectronic components like field-programmable gate arrays, application-specific integrated circuits and printed circuit boards (PCBs) are designed in the U.S. but are fabricated, tested and packaged in other countries.
Why the Latest Tools Aren’t Enough
This presents every facet of U.S. Armed Forces with a plethora of trust and assurance challenges because very real threats are posed by malicious or counterfeit chips from foreign countries run by governments that do not have the U.S.’s best interests in mind. Without the proper safeguards, it’s all too easy for poor-quality, fake or pernicious devices to be introduced into U.S. defense systems.
There are, of course, cutting-edge tools designed to scan microelectronics devices for issues, but these alone are not enough to ensure their safe and reliable operation on the critical platforms used by the U.S. Air Force and our nation’s other military sectors. Not only does the current breakneck pace of technological change make it impractical and impossible to constantly buy new products and implement wholesale changes, but in today’s world, new threats arise on a daily basis.
Given all these factors, the only truly cost-effective, dependable way to maintain safety and trust assurance is to employ enduring solutions that have the ability to continually learn and improve. This means that artificial intelligence (AI) and machine learning cannot just be buzzwords — they must be utilized to their full extent.
Lasting Solutions That Continually Improve Over Time
For example, there is a great need for tools that can verify and validate the trust and assurance of PCBs — the foundational building blocks or “brains” of most electronic devices — while scanning for counterfeits. Electronic systems have become incredibly complex and expensive and getting a single PCB ready for use can require 150 or more intricate steps. These crucial components are being used by organizations responsible for national defense and homeland security, but with the private commercial sector heading the modern-day microelectronics charge, the U.S. military has no choice but to rely on consumer off-the-shelf (COTS) PCBs from outside sources. As a result, cost and integration with current systems cannot be overlooked.
By leveraging the power of AI and machine learning, we can create and maintain a database of the “gold standard” of PCBs, ensuring that all scans are assessing authenticity using the most accurate and up-to-date information available. My company, Battelle, has developed a software suite called Rapid Assembly Inspection for COTS Security (RAICS) that analyzes PCB assemblies in this manner. Using complementary imaging hardware, RAICS generates multimodal signatures from an array of industrial sensors and quickly assesses the integrity of the COTS device being scanned. And thanks to AI and automated machine learning, the more RAICS runs, the more robust and adept it becomes.
Similar enduring solutions are needed to protect the embedded code that runs on communications buses within larger systems. The code and components of distributed control networks like smart grids, autonomous vehicle systems and medical monitoring are prime targets for hackers, but an algorithm that leverages machine learning can be applied to continually improve their ability to protect themselves from cyberattacks.
Battelle’s CANProtect technology attaches to and observes the commands sent through cyber-physical networks and searches for anomalies using explicitly programmed machine learning techniques. Like RAICS, the longer it’s in the field, the more adept it becomes at protection. CANProtect can also be used across services and installations, from Apache helicopters to aircraft refuelers.
Those of us who serve our country through the U.S. armed services know all too well that there will always be looming threats to keep at bay and the cybersecurity of our critical systems will continue to face increasing pressure and risk. While the latest shiny new tech objects may seem like the best defense, they can’t be the only tools in our arsenal. To effectively combat dangers and risks while working within the realities and confines of current budget and implementation constraints, it’s imperative that we focus on and emphasize enduring solutions that continually grow smarter over time. The trust and assurance of our nation’s defense systems merit airtight protection that can go the distance.
Learn more about Battelle’s capabilities in Microelectronics Trust & Assurance.
George Lecakes is the Vice President and General Manager of National Security at Battelle.