Every available drone collision-avoidance technology carries a structural tradeoff, and that gap — not any single sensor — now stands between the industry and routine beyond visual line of sight (BVLOS) operations. That was the verdict delivered by FAA, NASA and uAvionix representatives at AUVSI's Drone and AAM Policy Symposium, held November 9 and 10 at Nationals Park in Washington, DC.

The panel, moderated by Dallas Brooks of Wing, assessed the current state of detect-and-avoid (DAA) technology as the industry waits for the FAA's proposed Part 108 rule on BVLOS flight.

Why can't one sensor do the job?

Christian Ramsey, Chief Commercial Officer at uAvionix, walked through the tradeoffs that shape every technology choice operators and manufacturers face:

• Radar measures distance to other aircraft directly, but adds weight to the airframe and carries limits in range and field of view.
• Cameras deliver detailed visual information yet struggle to determine distance and to process data fast enough.
• Cooperative systems — transponders and ADS-B — let equipped aircraft share position electronically. They are relatively inexpensive, lightweight and battery-powered, which could make them practical for a wider range of aircraft. Their weakness: they only detect aircraft that carry compatible equipment.

"There is no perfect sensor," Ramsey said.

Dr. Chester Dolph, Research Lead for Intelligent Low Altitude Sensing at NASA, pointed to computer vision and machine learning as the path to better camera-based systems — improving performance while shrinking size and weight. He flagged a harder problem for developers: building sensing systems that work across different environments and aircraft types. Combining ground-based and airborne sensors, he suggested, may open opportunities to improve detection.

The panel drew one concrete comparison to explain why solutions diverge: a small delivery drone faces weight and power constraints that a larger commercial aircraft simply does not. Different operations will require different answers.

What is holding up Part 108?

Marcus Cunningham, Manager of the FAA's Technologies and Standards Branch, could not discuss the substance of the proposed Part 108 rule, which remains under review. FAA officials are barred from commenting on its details before publication.

"I can tell you that we've never been closer [to completion of the rule]," Cunningham said.

Part 108 aims to establish a regulatory framework for routine BVLOS operations and reduce the industry's reliance on individual waivers and approvals. No publication date has been announced.

Today the FAA evaluates detect-and-avoid capabilities case by case, as part of individual BVLOS approvals. Cunningham explained that these assessments weigh both the technology itself and surrounding safety measures: operations in remote areas present different risks than flights in congested airspace. That case-by-case method keeps approvals complex for applicants.

What does this mean for manufacturers?

Brooks identified a circular problem for technology developers. Manufacturers need clear standards to guide system development, but those standards must evolve alongside the regulations that define them. Regulators want evidence that a system meets safety requirements; manufacturers want defined requirements before they can develop and validate systems to produce that evidence.

The panel's conclusion for practitioners: collision avoidance is not a sensor-shopping exercise. A reliable system must account for aircraft size, operating conditions, equipment costs, and the ability to detect both cooperative and non-cooperative aircraft.

Advances in sensing technology, paired with clearer performance standards, could make these systems more practical and affordable. Until then, the industry's task is translating those advances into solutions that operators can deploy consistently across different missions and environments — while waiting for Washington to publish the rule that will define the market.