2024 BrainChip Discussion, page-4677

In a nutshell...

What is COGNITIVE radio?

A cognitive radio is a radio that can be programmed and configured dynamically to use the best channels in its vicinity to avoid user interference and congestion.

The cognitive engine configures radio-system parameters. These parameters include "waveform, protocol, operating frequency, and networking". This functions autonomously, exchanging environment information with the networks it accesses and other cognitive radios (CRs) and adjusts the radio's settings to maintain signal delivery.

What's it the most important application?

Efficient use of radio spectrum.

Radio spectrum is becoming saturated (many uses - TV, telecoms, SCADA, radar, emergency channels, beacons, ham radio etc.) the traditional approach is to carve-up the spectrum and allocate it to dedicated purposes and sell specific frequencies via licensing to avoid users crashing into each other.

This is analogous to air traffic control. Specific corridors through the navigable airspace are assigned to get traffic from one controlled airspace to the next. Aircraft traveling between cities are not visible to air traffic control (outside radar coverage) so must stick to their assigned corridor to avoid midair collisions.

However, as air traffic increases, these assigned corridors are becoming saturated. The solution being proposed (and trialed) is to install navigation equipment in each aircraft that is mesh networked allowing every aircraft to know exactly where other aircraft are located in real-time and for the system to use algorithms to guide all aircraft to their destination via the most fuel efficient route while constrained by the parameters of destination, schedule, collision avoidance (safe separation both horizontal and vertical), and predicted congestion at their destination airports.

Assigned spectrum is far from being fully utilized, and efficient spectrum use is a growing concern; CR offers a solution to this problem. A CR can intelligently detect whether any portion of the spectrum is in use, and can temporarily use it without interfering with the transmissions of other users. In this way (like the air traffic example) vastly more traffic can be handled by the available spectrum.

Put simply. If I have paid for exclusive use of 96.3 MHz but only transmit on it twice a day, that's not efficient use of spectrum. But, if my license allows my Cognitive Radio to use 50 MHz to 120 MHz then a lot more users can fit on that spectrum.

Why are the leading players in the Cog Radio mostly military primes?

Cog Radios support sophisticated frequency hopping "on the fly" allowing encrypted packet data to be transmitted securely but also jumping around to find clear spectrum. Particularly important when the enemy is using jamming.

However, it gets a lot more sophisticated including using Cog Radio to locate targets and a whole bunch of uses.

Such systems are the foundation for command and control systems that enable aircraft, ships, and guided weapons to share radar tracks and commands to provide maximum battle zone awareness, target identification, confidence levels, and communicate responses.

In the emerging field of drone swarming such systems will become critical.





Further sophisticated uses

As objects (aircraft, ships, drones, missiles for example) move through electromagnetic waves they cause detectable disturbances. In a hot zone (battle in progress) this electromagnetic radiation will be increased by jamming, radar and communications.

Cognitive Radio using high speed AI edge devices can exploit this saturated and chaotic (noisy) electromagnetic environment converting the disturbances caused by moving objects into images. This is a form of passive radar. Passive means that you can detect and identify without emitting your own radar signal.

While this sounds secret squirrel, researchers in Australia are working on a flood detection system that uses cell tower radio propagation that monitors vast areas of land looking for rising river levels and flood planes. Current resolution is less than an inch.

See GPU-Accelerated Signal Processing for Passive Bistatic Radar
https://www.mdpi.com/2072-4292/15/22/5421