The hardest problem in beyond-visual-line-of-sight flight has never been the flying. It's the seeing. A drone can hold a course, follow a waypoint chain, and land itself on a pad with more precision than most human pilots. What it cannot easily do is notice the crop duster coming out of the sun.

On July 13, 2026, two Israeli companies announced a joint field program that answers that question by refusing its premise. Instead of making the aircraft see, High Lander and Thirdeye Systems are making the ground see for it — and then telling the aircraft what to do about it.

Moving the sensor off the airframe

The architecture is the story here. Conventional detect-and-avoid (DAA) thinking puts the sensor where the pilot's eyes would be: onboard. That means every airframe carries the size, weight, power, and cost of its own radar or camera suite, and every airframe pays that penalty again on the next unit. For a fleet operator, DAA scales linearly with the number of aircraft — which is precisely the wrong curve for a business model built on flying a lot of cheap drones.

Thirdeye's contribution is MeduzaX, a passive optical radar that the companies say uses edge computing to process environmental data without latency. High Lander's contribution is Vega, the UTM platform that consumes those detections and turns them into a traffic picture. Per the announcement, the MeduzaX feed integrates directly into Vega, which aggregates inputs from detection and counter-UAS technologies into a single command-and-control view, cross-referenced against flight plans managed by High Lander's Orion drone fleet management platform.

The output is deliberately legible: authorized cooperative traffic renders as blue targets, uncoordinated or unauthorized aircraft as red. That's an IFF function in everything but name — friend or foe, rendered for a controller rather than a fighter pilot. The distinction matters operationally, because the red targets are the ones that constitute the actual regulatory problem.

What "passive optical" buys you

The word doing the most work in "passive optical radar" is passive. An active radar illuminates the sky and listens for returns. An RF-based detection system listens for a drone's own control links. Both approaches have well-understood tradeoffs, and a passive optical system sidesteps a specific cluster of them: no emissions, no spectrum license to obtain, and no signature to jam or home in on. A system that doesn't transmit is a system that doesn't announce itself.

Nor does it require the target to cooperate. The companies note the system operates without reliance on aircraft transponders. RF detection finds drones that are talking. ADS-B finds aircraft that are broadcasting. Optical detection finds whatever is physically in the frame — which, in the DAA context, is the entire point.

The tradeoff, of course, is that optical systems live and die by conditions, and the trial program appears to have been designed with that criticism in mind. The tests were run against aircraft above and below the horizon, against open-sky and terrain backgrounds — one clean, one visually cluttered — with approaches from the direction of the sun and varied lighting, including a controlled approach scenario between a helicopter and drones. Those aren't flattering conditions chosen to produce a good demo reel. Sun-line approach is the canonical failure mode for any camera-based tracker, and below-horizon detection against ground clutter is the canonical hard case. Choosing to publicize those specific test cases is a claim about robustness, and a reasonably falsifiable one.

The trial partners reinforce the point. Dronery operated the drones. Brook Aviation operated the helicopter. Helicopters are not an incidental choice — low, slow, often non-cooperative, and operating in exactly the altitude band where BVLOS drone operations want to live. MeduzaX is also described as operational with the Israeli Defense Forces, which is a different validation environment from civil aviation but not an irrelevant one.

Why It Matters

Strip away the vendor framing and this is a bet about where the industry's cost curve bends. Ground-based DAA trades per-airframe SWaP and cost for fixed infrastructure and coverage volume. You buy the sensor once, you site it once, and every aircraft that flies through its coverage gets DAA for free. The economics invert: instead of DAA cost scaling with fleet size, it scales with geography. For a delivery operator saturating a metro area, that's an enormous difference. For an operator flying two aircraft across a thousand kilometers of pipeline, it may be a worse deal.

The regulatory stakes are more concrete. The FAA's BVLOS NPRM — the proposed Part 108, published August 2025 and still not final as of this writing — would require unmanned aircraft operating in higher-risk environments to detect and avoid non-cooperative aircraft: aircraft not actively broadcasting their location via ADS-B Out or an approved alternate electronic conspicuity device. Proposed §§ 108.180(b) and 108.185(d)(5)(ii) attach that requirement to operations in Class B and Class C airspace and over Category 5 population areas. That single clause is the commercial chokepoint of the entire American BVLOS industry. Cooperative traffic is a solved problem; you receive a broadcast and you deconflict. Non-cooperative traffic — the helicopter, the general aviation aircraft, the ultralight with no transponder — is what nobody has cheaply solved at scale.

The NPRM's right-of-way construct sharpens the point further. Under the proposal, Part 108 UAS would hold presumptive right-of-way over crewed aircraft — except when the crewed aircraft is broadcasting its position via ADS-B Out or an approved electronic conspicuity device, is in Class B or C airspace, is departing from or arriving at an airport or heliport, or is over Category 5 population areas. Read that list carefully and it describes a regime where the drone's legal privilege evaporates precisely in the situations where a human is most likely to be somewhere unexpected. The exceptions are where DAA has to work.

A ground-based sensor that reliably picks up non-cooperative traffic and pushes it into a UTM is, in that light, not a nice-to-have. It's the thing the rule is asking for.

Not a sandbox

What separates this from the long parade of DAA demonstrations is the substrate it's running on. High Lander, founded in 2018, does not operate a research UTM. The Civil Aviation Authority of Israel granted Vega a first-of-its-kind air traffic management unit license, and Israel was the first jurisdiction in the world to make UTM connection mandatory for drone operations. CAAI regulation 10916, published November 23, 2023, requires drones weighing 200 grams or more to connect to an authorized UTM network and broadcast serial number, time stamps, location, altitude, velocity, and direction — adhering to the ASTM F3411-22a standard. High Lander manages Israel's mandatory nationwide UTM with Vega.

That means every legal drone in Israeli low airspace is, by regulation, a cooperative target. It also means this DAA trial is being layered onto real regulated traffic rather than a cordoned-off test range. The blue targets on the Vega display aren't simulated; they're the national picture. The red ones are what the optical layer adds.

There's a second wrinkle. Vega aggregates not just detection data but counter-UAS technology inputs into a single command-and-control view. That quietly erases the traditional boundary between air traffic management and airspace defense. A system that renders unauthorized aircraft in red and also ingests counter-UAS feeds is not merely advising pilots. It's a targeting-adjacent picture wearing an ATM interface. Whether that architecture travels to jurisdictions with different threat models and different civil-liberties instincts is an open question that the July 13 announcement does not address.

The humans still in the room

One nuance is worth holding onto against the autonomy rhetoric. Vega supports BVLOS operations through a staffed flight center operated by certified drone professionals — the arrangement High Lander describes at Israel's Ta'anakh solar project, where CAAI-approved Vega manages the airspace and certified BVLOS operators staff a manned flight center. There are people in a control room. The word "autonomous" in this stack describes the aircraft, not the system.

That's not a criticism so much as a description of what the DAA layer is actually for. Today, a human in the flight center is doing cognitive work that scales badly — watching, judging, intervening. A detection layer that reliably surfaces non-cooperative traffic and cross-references it against filed plans is what changes the operator-to-aircraft ratio from something like one-to-few to something like one-to-many. The DAA layer isn't replacing the humans. It's what lets fewer of them cover more sky.

"BVLOS operations are the future of aviation, opening up new possibilities for business, public safety, security, and defense," said High Lander CEO Alon Abelson in the announcement. "Getting there requires more than just tracking; it requires total airspace clarity." That's the boilerplate. The interesting sentence is the unspoken one underneath: the future arrives when somebody proves they can see a helicopter coming out of the sun, and prove it to a regulator, and do it without bolting a radar to every airframe. This program is an argument that the proof comes from the ground up.

Whether the FAA agrees is a separate question, and one the agency has not yet answered.

Sources