Thanks to recent technological advances, militaries can co-locate sensors and ruggedized, AI-powered supercomputers in tiny spaces across big networks – and acquire the information advantage they crave.
In October 2020, the U.S. Army and U.S. Air Force announced the Combined Joint All-Domain Command and Control (CJADC2) Implementation Plan, a unified effort to provide strategic advantage in large-scale, multidomain battles. Deemed a warfighting necessity by the U.S. Department of Defense, the CJADC2 framework helps inform modern warfare initiatives by better managing data volume and complexity. This initiative is designed to give U.S. military personnel a tactical advantage and provide the same across and within allied and partner nations, promoting a more collaborative and effective multinational defense strategy.
CJADC2 aims to create a multinational framework for integrated command and control that enables military forces to sense, make sense and act, in the DOD’s words, “at all levels and phases of war, across all domains, and with partners, to deliver information advantage at the speed of relevance.” In other words, it gives them decision superiority: the ability to assimilate, analyze and respond to information acquired from the battlespace more rapidly than an adversary.
While this definition captures what CJADC2 aims to achieve, it says little about how to achieve it. However, some lessons learned in recent conflicts have been integrated with this data-driven warfighting concept.
Growing Use of Digitized Information
CJADC2 uses artificial intelligence (AI), machine learning (ML), decision autonomy and other advanced capabilities to better connect sensors with shooters (e.g., soldiers, tanks, UAVs) and reduce the time it takes to bring lethal and non-lethal effects against an adversary to influence multidomain operations. Not surprisingly, AI-based, compact supercomputers designed to manage a growing amount of data and inputs are increasingly used in military and defense operations. By using this compact, rugged computing technology incorporated directly into today’s defense platforms, military operations gain better intelligence, faster, which leads to more successful outcomes.
Throughout history, decision superiority has always been crucial to winning or losing battles, with success destined for those who can best leverage and secure information to make the best decision in the shortest time. Military conflicts in the 21st century will continue to utilize this strategy at an accelerated pace, thanks to advancements in AI and data processing.
Objectives such as lowering the cognitive load of soldiers and decision-makers and decreasing the response time to gain an advantage represent just some of the requirements, risks and technical challenges being addressed.
The Importance of Shared Intelligence
Advances in telecommunications, sensors, processing power and weapons, along with the growing utility of space and cyberspace as operational domains, have fundamentally shifted the character of command and control in warfare. Data is the new strategic asset that is employed enterprise-wide in multidomain operations to achieve a holistic approach.
The benefits of networked communication include:
- Streamlining of large data transfers from sensors to mission computer
- Improved system response time
- Reduced wiring complexity increases system reliability, availability, maintainability
- Improved upgradeability and scalability
One iteration of CJADC2 focuses on creating a global targeting system that can enable combatants to locate, target and engage the enemy, then asses the results – a critical process known as the kill chain. Another looks at how CJADC2 can assist with achieving decision superiority to maneuver forces to positions of advantage to prevent an adversary from meeting their objectives. This iteration has recently been analyzed in manned and unmanned ground systems in land operations.
AI in Ground Operations
One of the most complex ground-based maneuvers is a wet-gap crossing. However, there are distinct logistical challenges in planning and executing these critical operations. When successfully executed, a wet-gap crossing operation can provide one of the most valuable assets in war – speed – to seize the initiative, prevent enemy reconnaissance, and exploit success. Executing a safe and efficient wet-gap crossing allows friendly forces to set the necessary conditions for further success.
A recent analysis of a failed wet-gap crossing by Russian forces in eastern Ukraine over the Siverskyi Donets River highlighted many challenges and risks associated with this complex operation and identified potential technical solutions using AI/ML and other critical technologies.
Information Flow Improves Risk Analysis
Since most future breaching operations will likely be conducted using unmanned or optionally manned systems, large amounts of data must be secured and transmitted across tactical networks to synchronize reconnaissance and security, logistics and other warfighting functions.
At the macro level, CJADC2 involves gathering massive quantities of data through a broad range of distributed sensors and processing it into actionable information. The system is stitched together with a robust set of communication links that allocate the correct information across the network to enable organizations to achieve enhanced effects in their specific areas of responsibility.
The OODA Loop, a well-known and accepted decision-making model, describes a four-step process for executing combat operations: Observe, Orient, Decide, Act. Developed by U.S. Air Force Col. John Boyd, it emphasizes the importance of speed and agility in decision-making and action-taking to complete the loop as quickly and efficiently as possible so that you can adapt to changing circumstances and take advantage of opportunities as they arise.
Deploying AI algorithms on devices physically close to the data source – an approach known as “AI at the edge” (AIAE) – allows decisions to be made with minimal latency and provides flexibility in a rapidly changing environment. For example, connecting sensors directly to the AIAE unit will greatly reduce latency between the observe and orient steps in the OODA Loop.
It is the same for significantly reduced latency between the orient and decide steps because there’s no need to send out large amounts of data for additional decision-making steps to a distant node and then wait for the decision to be sent back. Sending the resulting “act” command from the AIAE unit reduces latency for the decide-act steps for the same reasons.
Decision superiority through processor design
A dominant commercial-off-the-shelf solution for AIAE processing is a general-purpose graphics processing unit (GPGPU). They can handle large amounts of data in parallel – much faster than traditional central processing units (CPUs) – thereby accelerating a wide range of AI applications.
Modules in the NVIDIA Jetson family combine AI-capable GPGPUs with multicore CPUs to create a tightly coupled, high-performance, low-power supercomputer that supports AI processing and decision-making applications software.
For example, the NVIDIA Jetson Xavier NX module provides six trillion floating point operations per second (TFLOPS) performance with a maximum power of 15 watts. This performance is comparable to that of a several hundred-watt workstation with a processor and GPU cards.
This type of computing architecture can process and apply AI algorithms for more than 20 high-definition video inputs with 1040p resolution at a rate of 30 frames per second – enough bandwidth to run AI applications for a system of multiple high-definition cameras. For defense operations, the high processing capabilities of the NVIDIA architecture enable AIAE processing, thanks to the compact supercomputers embedded within the military platform.
A ruggedized supercomputer with an NVIDIA Jetson Xavier NX module can be as small as 4” x 2.3” x 3.9”. With its low power consumption and maximum weight of 1.3 lbs., it’s an ideal candidate for AIAE applications from the perspectives of performance and SWaP: size, weight and power.
Other Considerations of AI in Military Operations
AIAE’s numerous benefits, such as reduced latency and increased security, also present some technological challenges, including limited processing power and storage as well as energy efficiency.
Using rugged AI supercomputing modules addresses many of these challenges, but there are also concerns of data transfer and security.
Time-sensitive networking (TSN) is a communication protocol that ensures critical information reaches decision-makers without delay by transmitting real-time data with high precision and reliability. It also facilitates the collection, aggregation, and analysis of this real-time data, empowering decision-makers with accurate, up-to-date information.
TSN synchronizes devices and systems across distributed networks to ensure that data from multiple sources is aligned and consistent. This provides a holistic view of the operational environment and enhances coordination between different components, such as sensors, actuators, and control systems, for seamless collaboration and integration.
This brings into play AIAE’s cybersecurity parameters to ensure high-performance AI-capable systems are protected from cyber and spoofing attacks, securing shared information in several ways. These include reducing the amount of data shared across tactical networks, simplifying data distribution efforts, reducing system latency, improving data redundancy at the sensors and eliminating interoperability issues between systems since all use the same communications protocols and data messaging structures.
Leveraging AI/ML and advanced algorithmic warfare systems provides a significant decision-making advantage. Rugged, compact supercomputers can help manage the influx of data that systems must handle while providing improved intelligence in military operations.
Every system and program should mandate sensor data sharing and interoperability. This data-sharing construct can create secured battlespace awareness, in which actions in one part of the single, integrated, global battlespace can be understood and informed by actions and decisions required in other areas.
Timothy Stewart, BSME, is Director of Business Development at Aitech, which develops rugged embedded-computer solutions for industrial, military and aerospace applications. He has 20 years of experience in high-technology hardware, software and networking products. Timothy holds a BS in mechanical engineering and physics from Boston University.
