Harnessing big data: How advanced analytics and AI are changing the battlefield

Story

August 05, 2024

Today’s warfighters are inundated with data from numerous sources, a stark contrast to the relatively limited data available during the wars in Iraq and Afghanistan two decades ago. During past conflicts, soldiers collected data primarily through traditional means such as reconnaissance reports, human intelligence, and satellite imagery. While valuable, these sources provided limited information compared to the multifaceted data streams available today.

In-theater personnel generate data from a wide array of sources, including advanced sensors, uncrewed systems, real-time communications, and satellite feeds. The sheer volume of data collected is staggering. For instance, sensors on aircraft and drones can generate terabytes of data during just a single mission. This data includes high-resolution imagery, radar, electronic warfare (EW) information, and more.

There are a host of challenges when it comes to managing all this data and making it useful to the warfighter, says David Mercado, director of field engineering at Wind River (Alameda, California). These hurdles include “shortening the OODA [observation, orientation, decisions, and action] loop to accelerate and optimize battlefield decision-making (i.e., speed and accuracy); collecting data from all relevant sources (which may be overwhelming in terms of volume and speed) such as sensors or devices, making sense of that data in situ and in the context of data from other sources, and analyzing the data to understand the relevant possible courses of action.”

The visualization of this data has also evolved: Previously, data would be processed and analyzed by a limited number of analysts and then disseminated in reports. Today, data is often visualized in real time through advanced software platforms, providing commanders and soldiers with immediate insights and actionable intelligence.

Big data needs cutting-edge tech

Ideally, this data should boost situational awareness, inform strategic and tactical decisions, and predict future threats.

For example, sensor data collected from drones and surveillance systems can provide detailed imagery and movement patterns of enemy forces. Communication intercepts can reveal plans and intentions, while logistical data ensures that resources are deployed efficiently. This holistic view of the battlefield allows commanders to make informed decisions that enhance operational effectiveness and reduce risks.

The problem is that the utility of this data is heavily dependent on technology: Can all this data, as useful as it is, be processed and delivered to warfighters in a timely manner? If not, it does not do anyone much good.

This is where advanced technologies such as artificial intelligence (AI) and machine learning (ML) come into play. These technologies can analyze vast datasets quickly and accurately, identifying patterns and trends that would be impossible for human analysts to discern.

“By sorting and deciphering data using learned patterns programmed into it, AI can fill in gaps in data (infer), make decisions, and recommend or even take actions faster than a human operator could,” Mercado says. “Over time, as the AI engine is exposed to more data and patterns, it will also possess the ability to recognize trends in the data, predict outcomes, and adapt accordingly.”

For example, Wind River’s VxWorks integrates AI and ML frameworks to optimize embedded systems for real-time data processing. This enables systems to analyze and act on data with minimal latency, the company says. (Figure 1.)

[Figure 1 | Wind River’s VxWorks is a real-time operating system (RTOS) that provides a scalable and secure environment for mission-critical computing systems.]

This challenge can be addressed with technology that can process significant amounts of sensor data at the edge through AI and ML, with the goal of ensuring that only the most relevant and critical data is distributed across the network, according to an Abaco Systems spokesperson. The idea is to not only conserve bandwidth, but also enable faster decision-making across domains for the Joint All-Domain Command and Control (CJADC2) framework, they say.

For these applications Abaco offers the SBC3901, a 3U VPX single-board computer (SBC), to process complex AI and ML tasks in harsh environments in real time. (Figure 2.)

[Figure 2 | Abaco Systems’ SBC3901 single-board computer.]

Predictive analytics

Predictive analytics is another promising application of big data in the military. By analyzing historical data and identifying patterns, predictive analytics can forecast future events and trends, enabling senior leadership to anticipate and prepare for potential threats – for instance, analyzing past supply-chain data can help predict future logistical needs.

By monitoring data such as fuel consumption, maintenance records, and deployment history, forces can anticipate equipment failures and perform maintenance proactively, thereby slashing downtime and extending an asset’s life cycle.

The integration of real-time data analytics also helps with decision-making on the battlefield. Commanders can receive up-to-the-minute information about the status of their forces, enemy movements, and environmental conditions. This real-time data allows for rapid adjustments to strategies, troop movements, and tactics, increasing the likelihood of mission success.

Challenges in accessing data

The military struggles with bandwidth limitations. The world is saturated with sensors, communications devices, and unmanned systems, all generating vast amounts of data. Transmitting this data in real time to command centers or data processing facilities can strain existing communication networks, leading to latency issues and potential data loss. All of these moves are even more difficult in remote or hostile environments where infrastructure may be limited, compromised, or destroyed.

“Battlefields have difficulties with band­width limitations, communication latency, interoperability across the services and data security,” Mercado says. “Today’s broad data-collection systems provide access to an overwhelming amount of information that needs to be correlated and assessed.”

Another factor: the security of data transmission and storage. Secure data encryption and robust cybersecurity measures are essential, but these can add layers of complexity and slow down data access and processing speeds.

The growing importance of real-time data visualization

Collecting and storing data is only the beginning. Making this data useful requires overcoming several additional challenges. One of the most significant is the ability to visualize it for the warfighter.

Integrating this disparate data to form a coherent picture is not an easy task, and one that requires sophisticated algorithms and advanced analytical tools. But it’s an important puzzle to solve, Mercado says. “The point of data in battle is to enhance combat effectiveness – data at speed and scale – to make use of data faster than adversaries,” he says.

Wind River’s VxWorks seeks to tackle this challenge by providing a high-reliability real-time operating system (RTOS) that gathers data from various sources, interprets that data, and provides it to the war­fighter at the edge.