Embedded Systems and AI Give the Defense Sector the Edge

 

The networked world is gaining strength at the edges. Computers are becoming smaller, cheaper, and more powerful. Sensor technology is evolving rapidly, and enhanced connectivity is enabling lightning-fast data exchange. All this opens up possibilities that were unthinkable outside big tech research centers just a few years ago.

Edge AI has arrived, and the combination of embedded devices and machine learning algorithms creates immense new opportunities to transform ordinary military personnel's activities. 

A soldier-borne compute module weighing 400 grams is so powerful that it can process images from up to five cameras simultaneously. It can also integrate input from optoelectronic devices such as night vision, thermal imaging systems, head-up displays, and augmented reality glasses. Additionally, it processes data from vibrational sensors and microphones. 

A soldier-borne radio communication module and a battery pack become a sophisticated, two-legged node in the military network, harnessing AI to optimize its mission and reduce risk. 

Interlinked System

Edge AI in defense is not limited to the individual. Equipment and assets of any size and type, from ground vehicles to sea vessels and aircraft, use Edge computing. The most advanced combat aircraft in existence, the F-35, can fuse sensor data with other aircraft and share a single picture of data across jets flying in formation. This interlinked system provides a more thorough assessment of threats and targets than a single aircraft could provide on its own.

The Digital Soldier concept relies on the ability to process data where it is captured. Being independent from a centralized computing system saves valuable time and can make a crucial difference in the heat of the moment.

AI solutions must have several components to work efficiently. Edge devices such as drones, vehicles, or wearable equipment must be able to process the data they receive, sharing the tasks with the edge servers. The Edge devices must also be equipped with enough computing power to handle multiple devices. 

The AI models running on the equipment must be optimized for this Edge infrastructure, which has obvious limitations in memory and power consumption. Strong connectivity is required, as many Edge AI applications, although able to function in air-gapped environments, use fast data transfer when available. Robust cybersecurity is a crucial prerequisite across the entire system. 

Multi-Party Collaboration

Building such a complex and multi-layered Edge AI solution often requires multi-party collaboration. Large and well-established military equipment manufacturers can cooperate with engineering development companies. These companies usually have expertise in adjacent areas, such as industrial automation, automotive, or oil and gas.

The Internet of Things (IoT) technologies developed for these domains provide a solid foundation that can be adapted to specific defense requirements. Smaller companies with niche expertise in machine learning or computer vision algorithms can contribute their expert knowledge to develop the defense-edge AI applications of tomorrow.

The expertise Akkodis brings is centered around connectivity and Edge computing. Its engineers have developed edge computing platforms designed to improve performance for applications such as industrial automation, autonomous driving, and Edge AI. These applications are designed to enable increased performance efficiency through additional GPU (Graphics Processing Unit), vision, motion, and I/O (Input/Output) cards. They come in ruggedized versions designed to cope with the extreme demands of defense. Ultra-rugged portable and vehicle-mounted nodes are used for tactical communication. Routers and tablets - for handling drones and cameras are built to survive the harshest environments. 

In defense, Edge AI is a bridge reaching all the way from command-and-control centers to frontline personnel – a short-cut from data to decision.

From data to Decision

Akkodis offers another piece of the Edge AI puzzle. Its software developers specialize in optimizing AI models to perform customer-defined tasks. The team helped a large home security camera manufacturer use AI to optimize its software and prevent false alarms.

Akkodis designed and implemented a decision support tool to improve delivery capacity allocation. An Akkodis client was experiencing stock management issues and needed a tool to visualize the stock and storage space situation. The Akkodis AI-powered solution included forecasts and prediction models highlighting upcoming problems. Both can be adapted to military surveillance or military logistics.

As the demand for defense technology grows and the technological potential of Edge AI increases, Akkodis is ready to contribute its expertise to innovative solutions that were unimaginable a few years ago. Those solutions harness the power of the network edge. AI at the data’s source is a powerful tool waiting to be developed and refined.

In defense, Edge AI is a bridge reaching all the way from command-and-control centers to frontline personnel – a short-cut from data to decision.