The Rise and Rise of Artificial Intelligence in Military Applications. TVS Diodes Offer Protection and High Performance to Mission Critical Hardware
Disrupt or be disrupted. It seems everywhere you turn, some industry or company has applied Artificial Intelligence technology to improve one aspect or another of their operations. In our data-rich and hyper-connected world, the fact should not surprise. And yet, much of that data remains untapped. AI can play a role in extracting insights and coming to the right conclusion, whenever a decision is to be made.
In the military where the stakes are high, the application of AI is also rapidly growing with implications on national security. Advanced nations around the world are working intensely to develop innovative AI applications for various functions that will give them the winning edge on any battlefield.
Incorporating AI into military applications relies on the use of advanced sensing instrumentation, complex computing, and reliable communication, all of which are built upon electronic systems. Unfortunately, these systems are vulnerable to power irregularities that can occur, namely Electrostatic discharges, ESDs, and induced transient surges. To overcome these problems, protection devices should feature to ensure smooth and reliable operations.
MDE Semiconductor, Inc. is at the forefront of innovation with a range of TVS diodes that clamp overvoltages to protect downstream loads from destructive transient events. Our 30KP series of TVS diodes can handle up to 30000 W transient power induced by ESD, load dump or lightning. Weighing just 0.07 oz (2.19 g) and capable of any mounting positions, the series of diodes offer circuit designers high flexibility to simplify their circuits for novel applications.
At the same time, excitement about AI’s potential uses still needs to be filtered through the lenses of scalability and high-value addition (when compared to current capabilities without AI) Therefore, in this article, we’ll be looking at what constitutes Artificial Intelligence, the US military’s incorporation of the technology, and key issues that remain a challenge to its adoption.
Defining Artificial Intelligence is a Challenge
The rise in AI R&D has been made possible by three key enablers, namely:
- Availability of sources of big data
- Improvements in machine learning capabilities
- Advancements in computer processing power
However, there is no one standard definition of Artificial Intelligence. The discrepancy is mainly because potential applications for AI are so broad, and various sections of the research community have used/ are using multiple approaches towards adopting it. To add to the confusion, during independent development, some terminologies are assumed which can often conflict in meaning from one community to another.
Section 238 of the National Defense Authorization Act for Fiscal Year 2019, includes the following categorization of AI as:
- An artificial system performing tasks under changing and unpredictable conditions without significant human intervention, or when exposed to data sets, can learn from experience and improve performance.
- An artificial system designed to perform tasks requiring human-like cognition, learning, perception, communication, planning, or physical action, whether developed in physical hardware, computer software or any other context.
- An artificial system featuring cognitive architectures and neural networks, designed to think or act like a human.
- A set of techniques developed to approximate cognitive tasks, such as machine learning.
- An artificial system designed to act rationally, including an embodied robot or intelligent software agent that performs tasks using perception, reasoning, learning, planning, communicating, decision making, and acting.
Following the categorization above, a few concepts need further elaboration (they invariably show up whenever AI in military applications is discussed):
Machine Learning: involves statistical algorithms that are trained on large data sets to derive procedures using analysis to solve the same category of tasks, but in a new situation (replicating human cognition).
Automated system: operates with no (or little) human involvement, and the system is limited to well-defined tasks that have predetermined outcomes.
Autonomy: involves self-governance for performing assigned tasks, and the system relies on its perception, analysis of a situation, planning, and decision making. For example,
- Lethal Autonomous Weapon System, LAWS, once activated will engage targets without further human intervention or operation
- Human-Supervised Autonomous Weapon System requires human oversight with the ability to terminate engagement
- Semi-Autonomous Weapon System is activated to engage specific targets that a human operator selects
Robot: executes a set of actions, whether directed by a human, a computer, or a combination of both.
The complex electrical and data networks which help AI-enabled systems operate can be vulnerable to interference from any connected electronic equipment. This means that transients caused by power anomalies can propagate throughout the connected circuits. TVS diodes are well-suited for ESD protection and DC power protection to ensure the reliability and performance of military equipment.
Applications of Artificial Intelligence in the US Military
The Department of Defense, understanding the need to strengthen US competitiveness has established several initiatives including the Joint Artificial Intelligence Center (JAIC). The objective is to coordinate efforts across various units (e.g., DARPA, IARPA) in over 600 active AI projects. Below are the ways Artificial Intelligence has been incorporated or is being tested in military functions:
With large fleets of aircraft, ground vehicles, sea vessels, and other mechanized equipment, the business of performing inspections to replace and maintain parts is not a trivial one. Instead of standardized maintenance schedules and reactive repairs when parts get damaged, the Air Force is testing the use of AI to create predictive maintenance schedules that are tailor-made to individual aircraft. The algorithm pulls data from real-time sensors in the onboard systems and engines of the aircraft to determine when best to replace or make repairs.
The Army’s Logistics Support Activity (LOGSA) contracted IBM’s Watson AI software in 2017 to optimize the Army’s shipping flows for the distribution of repair parts. The objective is to determine the most cost-effective and time-efficient way to deliver supplies, a job that used to be performed by human analysts.
Intelligence, Surveillance, and Reconnaissance
With advancements in surveillance technology, the amount of data generated is often insurmountable for human analysts to sift through to gain valuable insights. AI is excellent at automating such manual effort. Success was demonstrated in operations in Iraq when machine learning algorithms combed through several hours of aerial video footage to identify hostile activity promptly.
The threat of cyber attacks on US security and intelligence agencies is ever-present. The volume of activity required to detect, evaluate, and patch system vulnerabilities, when faced against highly incentivized malicious intentions, is simply staggering. Instead of relying only on human intelligence, not to mention endurance, AI-enabled cyber tools that can detect and defend against network anomalies in a few seconds is a more attractive approach.
Due to the increased vulnerability of miniature circuit components, threats of voltage transients and power surges are also an issue. Thankfully, there are tools available for overcoming these risks, namely, TVS diodes and other surge protection devices.
AI has made it possible for adversaries to create realistic image, video, and audio forgeries as part of information warfare. In doing so, they can influence public opinion and the actions of high-ranking officials. DARPA launched its AI project, MediFor, to overcome the shortcomings of current forensic analysis tools at detecting media manipulations and how they were performed.
Command and Control
To centralize the planning and execution of operations (sea, space, land, or air), the Airforce working with Lockheed Martin, and other startups are developing an AI-based system that creates a comprehensive picture from all available input data (Multi-Domain Command and Control, MDC2). The objective is to help stakeholders, who previously had to rely on multiple platforms providing sometimes contradictory information, make better decisions faster.
Semi-autonomous and Autonomous Vehicles
Receiving the bulk of general press attention is the AI-enablement of military vehicles (ground vehicles, drones, fighter aircraft, etc), probably due to parallel reports about smart vehicles for commercial use. AI tools would extract data about obstacles in the environment from sensors, navigational data from GPS and other data to plan and execute movement, and communicate with other vehicles.
The Air Force Loyal Wingman project is one such example with uninhabited old-generation fighter jets pairing with a manned F-35, such that a wingman performs tasks for its lead.
Swarming of enemy targets by smaller less expensive autonomous vehicles is another potential application of the technology.
Key Challenges Facing Artificial Intelligence for Military Applications
The large majority of projects for AI integration into military operations are still at the testing and development phase, and standing in the way of full-scale adoption include:
- Attitude by the military towards AI-enabled systems and operations
- Continued investment towards research
- Advancement of the capabilities of AI beyond its current vulnerabilities
- Willingness of commercial sector AI experts to collaborate on military applications
These issues notwithstanding, many experts agree that the future of warfare will involve Artificial Intelligence in some shape or form. The pressure to compete with rivals who may not have the same ethical, legal, or moral stance about the technology’s use is hard to ignore.
MDE Semiconductor, Inc. Manufactures the 30 KP Series of TVS Diodes to Protect AI-enabled Military Systems Against Transient Overvoltages
As AI-enabled military systems call for more advanced sensing, computer processing, network connectivity, and electro-mechanical capabilities, so do the number of electronic components increase. Every electrical circuit requires protection against imminent damage from power surges, electrostatic discharges (ESD), overvoltages, and other catastrophic events resulting from irregular flow of electric power.
Transient Voltage Suppression diodes (TVS diodes) are circuit protection devices placed intentionally to suppress damaging overvoltage spikes and promote controlled electric flow. MDE Semiconductor, Inc. offers an extensive portfolio of TVS diodes suitable for the high-performance needs of naval vessels, ground vehicles, and fighter aircraft. Our 30 KP series combines up to 30000 W of peak pulse power dissipation in a compact 0.07 oz package. Soldered according to MIL-STD-750 standards, it is designed to survive high temperature environments. In addition, it offers fast response times less than 1.0 ps which is very attractive for high reliability.
MDE Semiconductor, Inc. aims to keep innovating to provide the support that upcoming military applications require.