Neural Net: The Future of Cooperative Air Combat Networking

By the early 2030s, the battlespace had become a complex web of stealth aircraft, advanced electronic warfare systems, and multi-domain sensor fusion. Traditional cooperative radar techniques, while revolutionary in the early 2000s, were increasingly vulnerable to sophisticated jamming, signal interception, and pattern exploitation. The Swedish Air Force’s response to these emerging threats was Radar-Samverkan 2.0, codenamed “Neural Net” — a distributed, AI-driven, multi-platform combat network designed to remove single points of failure and extend the survivability of both manned and unmanned assets. Built around the Gripen E Block IV, Saab’s GlobalEye AEW&C, and MQ-28 Ghost Bat drones, Neural Net represented not just an upgrade, but a complete rethinking of how air forces cooperated in high-threat environments.

1. Scramble & Takeoff

A. F 21 Wing, Luleå — Swedish/NATO Perspective

Year 2032. Snow whipped across the hardened shelters as the alert horn blared. Lieutenant Sofia “Valkyrie” Lindström climbed into her Gripen E Block IV, the cockpit illuminated by the glow of the Wide Area Display (WAD). On startup, the Mission Computers pulled not just the mission data package from the MDC via MIL-STD-1553B and ARINC 429 buses — but also a live sensor fusion map from the orbiting Saab GlobalEye AEW&C and a pair of MQ-28 Ghost Bat loyal wingman drones.

Her PS-05/A Mk5 AESA radar ran self-check silently in standby, while the TIDLS Mk.2 directional beam link came online — multi-band, frequency-hopping, beam-steered, with anti-jam algorithms hardened against even DRFM-based ECM. Meteor Mk.II missiles were armed, IRIS-Ts were green, and her WAD displayed a real-time AI-generated emitter plan, assigning which platform would radiate based on geometry, threat priority, and RCS of approaching contacts.

One slot in the four-ship formation was filled by Ghost 34-Alpha, an optionally manned Gripen drone configured for high-power radar duty and sacrificial illumination if needed. Throttles to full afterburner, Sofia rocketed down Runway 14, climbing through the snow clouds into the thin Arctic sunlight.

B. Olenya Air Base, Kola Peninsula — Russian Perspective

At the same moment, four Su-57 Felons of the 485th Guards Fighter Regiment lifted off on afterburner. Colonel Alexei “Grom” Mirov led the formation westward, climbing to high transonic cruise at 45,000 ft. Their N036 Belka AESA radars and 101KS-V IRSTs were already scanning, but today’s mission called for silent approach until within weapon release range.

The ECM suite in each Felon was configured for adaptive interference, learning enemy emission schedules in real time and applying targeted jamming. Mirov’s plan was to bait the Gripens into predictable illuminator patterns and then saturate their data links with tailored noise.

2. First Contact — Multi-Sensor, Multi-Platform Fusion

A. Swedish/NATO

At 320 km, Sofia’s WAD populated the first merged tracks. She wasn’t painting them herself — the GlobalEye’s Erieye ER radar and the MQ-28’s IRST pods were feeding data into the TIDLS Mk.2 mesh. AI-based cross-platform fusion stitched radar, IRST, and passive RF detections into a single, precise target track, with position errors under 50 meters.

The Neural Net emission scheduler chose Ghost 34-Alpha to act as illuminator first, not a human pilot. The AI factored in the drone’s position off the Felons’ nose and its high RCS decoys — any jamming effort against it would not compromise the crewed Gripens.

B. Russian

Mirov’s passive sensors detected the drone’s AESA burst — narrow, agile, and low dwell. His ECM officer attempted a noise spike, but the beamformed TIDLS link used directional, pencil-thin RF lobes and frequency-hopped across bands faster than their suite could follow. The Felons still had no solid range data on the crewed Gripens, only a high-RCS decoy signal from the drone.

3. AI/ML-Based Emission Scheduling in Combat

A. Swedish/NATO

As the range closed to 200 km, the AI reassigned illumination to Ghost 32, which had shifted altitude for better radar horizon. The switch happened mid-burst — imperceptible to human pilots — and denied the Su-57s any pattern to exploit. When Himalayas ECM concentrated jamming energy toward Ghost 32, the AI instantly switched emission to Sofia’s Gripen, while the MQ-28 went fully passive.

The dynamic geometry-based illuminator handoff meant no single aircraft radiated long enough to be triangulated. Sofia’s WAD showed the Su-57s in bright, solid track boxes, each labeled with fused IRST and radar quality indicators.

B. Russian

Mirov’s formation tried to force a break in the network by executing a high-speed split — two Felons climbing, two descending. The AI compensated instantly, assigning the GlobalEye’s wide-area radar to maintain both groups and using the MQ-28’s IRST to track the low pair through thermal contrast.

4. Higher-Bandwidth, Anti-Jam Data Links in Action

A. Swedish/NATO

TIDLS Mk.2’s beyond-line-of-sight relay through the GlobalEye kept every node connected even during terrain masking passes. When the Su-57s activated a swept-noise DRFM barrage, the AI adjusted link modulation rates and hopped to an alternate Ka-band uplink, beam-steered through the drone’s high-gain antenna. Track quality never dipped below 98%.

Sofia authorized a cooperative passive targeting solution — combining her Skyward-G IRST angle data with the MQ-28’s EO/IR and the GlobalEye’s ESM bearings to generate firing-quality solutions without using any onboard radar.

B. Russian

Mirov’s RWR remained quiet — dangerously so. The Felons’ MAWS gave no warning; no continuous radar meant no conventional launch detection. His only clue was an unexplained uptick in datalink chatter between NATO nodes, but by the time he suspected a missile in flight, the Meteors were already inbound.

5. Integration with Next-Gen Weapons

A. Swedish/NATO

Sofia launched two Meteor Mk.II missiles at 120 km. Guidance updates came not just from her Gripen, but also from the MQ-28 and GlobalEye — a redundant tri-source mid-course guidance network. Even if one illuminator was lost, the others could keep feeding the missile.

The AI also coordinated networked missile swarming — her Meteors and Ghost 33’s Meteors shared in-flight data over a missile-to-missile link, automatically allocating which would pursue which target based on evasive maneuvers. At 20 km, both missiles went pitbull, their active seekers locking before the Felons had a chance to beam effectively. One Su-57 exploded in a fireball, another took heavy damage to its port engine.

B. Russian

Mirov’s wingman screamed a missile warning far too late. Flares and violent jinks came seconds before impact — the Meteor’s throttleable ducted rocket had plenty of energy for terminal lead pursuit. Mirov himself broke into afterburner, accelerating to Mach 1.8 in a shallow descent, the remaining Felon beside him. The damaged jet limped east, its stealth compromised by missing paneling.

6. Disengagement & Return

A. Swedish/NATO

With two Felons neutralized and the rest fleeing, the AI reconfigured the formation into a defensive egress pattern, keeping TIDLS Mk.2’s beams tight on the retreating threats until they dropped off the GlobalEye’s scope. Sofia’s WAD replay buffer had already logged every emission, every IRST track, and every missile telemetry packet for post-sortie analysis.

B. Russian

Mirov ordered an abort, furious at the complete denial of ECM effect. The squadron’s techs would have to study why their adaptive jamming suite failed against what looked like a non-deterministic illuminator pattern. The loss of a Felon and the severe damage to another in a BVR exchange would sting politically and tactically.

7. Post-Mission Debrief

A. NATO Debrief — Luleå

The After-Action Report noted 100% track continuity, zero emission pattern predictability, and successful BVR kills without any radar scheduling input from human pilots. The distributed sensor-web — GlobalEye, MQ-28, and Gripens — proved there was no single point of failure. The AI’s threat-aware emitter switching and multi-source missile guidance meant even heavy ECM could not degrade the kill chain.

B. Russian Debrief — Olenya

Mirov’s regiment concluded that the NATO formation had eliminated the classic vulnerabilities of cooperative radar tactics: no fixed illuminators, no exploitable patterns, and no reliance on a single data link. Recommendations included developing multi-band, beamformed ECM, deploying counter-drone interceptors, and using long-range IRST cueing to detect passive network builds before missiles were in the air.

8. Conclusion

Neural Net demonstrated that the future of air combat would be won not by the most powerful radar or the fastest fighter alone, but by the resilience, adaptability, and intelligence of the network that connects them. By eliminating single points of failure, integrating diverse platforms, and leveraging AI for both emissions control and targeting, Radar-Samverkan 2.0 allowed small numbers of fighters to dominate the battlespace against numerically superior and technologically advanced foes. In an age where stealth and ECM were designed to blind traditional sensors, Neural Net turned the tables — making the enemy visible without ever revealing itself. It was not just an upgrade to Gripen’s cooperative tactics; it was the blueprint for the 21st century’s distributed aerial warfare. 

Note: This story is entirely fictional and does not reflect any real-life events, military operations, or policies. It is a work of creative imagination, crafted solely for the purpose of entertainment engagement. All details and events depicted in this narrative are based on fictional scenarios and have been inspired by open-source, publicly available media. This content is not intended to represent any actual occurrences and is not meant to cause harm or disruption.