When a manned aircraft pilot flies in uncontrolled airspace, one of their core responsibilities is to see and avoid other aircraft. This is a fundamental principle of aviation safety — the pilot’s eyes are the primary mechanism for detecting traffic conflicts and taking evasive action. When an unmanned aircraft operates beyond visual line of sight, there is no pilot in the aircraft to perform this function. Something else must do it.

That something else is a detect and avoid system — sometimes called DAA, and sometimes referred to as sense and avoid or, in older literature, see and avoid. Whatever the terminology, the function is the same: the unmanned aircraft must be able to detect other aircraft in its vicinity, assess whether a collision risk exists, and take appropriate action to avoid it. For BVLOS operations in shared airspace, DAA capability is not optional; it is a foundational safety requirement.

Cooperative and non-cooperative traffic

The detect and avoid problem divides into two fundamentally different challenges depending on whether the traffic to be avoided is cooperative or non-cooperative.

Cooperative traffic is traffic that is actively broadcasting its position and identity — manned aircraft equipped with ADS-B transponders, other drones equipped with Remote ID, and aircraft operating under instrument flight rules with mandatory transponder equipment. For cooperative traffic, detection is relatively straightforward: the DAA system receives the broadcast, processes the position and track data, and assesses whether the reported position creates a conflict with the UAS’s planned route.

Non-cooperative traffic is traffic that is not broadcasting — recreational aircraft without transponders, gliders, paragliders, birds, and other objects that produce no electronic signature. Detecting non-cooperative traffic requires active sensing: the DAA system must detect the physical presence of the object using sensors that do not depend on the object transmitting anything.

Most real-world airspace contains both cooperative and non-cooperative traffic, and a complete DAA solution must address both. The relative proportion varies significantly by environment — urban airspace tends to have more cooperative transponder-equipped traffic; rural and low-altitude airspace tends to have more non-cooperative recreational and ultralight traffic, as well as bird activity.

ADS-B receivers

Automatic Dependent Surveillance-Broadcast is the standard technology through which equipped aircraft broadcast their identity, position, altitude and velocity. ADS-B In receivers — the receiving component, as opposed to the transmitting ADS-B Out component — allow a UAV to receive and process the position data broadcast by other ADS-B-equipped aircraft in the vicinity. This provides detection of cooperative manned traffic with relatively low cost and weight impact on the UAV.

The limitation of ADS-B for DAA is coverage: not all aircraft are required to be equipped with ADS-B Out, particularly at low altitudes where drone delivery operates. In the United States, ADS-B Out is required in most controlled airspace, but recreational aircraft operating under Part 107 or the recreational exception are not required to broadcast. In European airspace, requirements vary. The result is that ADS-B alone does not provide complete cooperative traffic coverage.

Radar and acoustic detection

For non-cooperative traffic detection, radar is the most capable technology available at the relevant scales. Radar systems can detect objects that produce no electronic signature, providing coverage that ADS-B cannot. The challenge for drone delivery applications is that radar systems capable of detecting small aircraft at relevant ranges have historically been too heavy and too power-hungry for integration into small delivery aircraft — though miniaturisation trends are progressively improving the trade-offs.

Acoustic detection — using microphones to detect the sound signatures of approaching aircraft — is a lower-cost and lower-power alternative that has been explored for some applications. Its effectiveness is limited by range (acoustic signatures are detectable only at relatively short distances), by wind and background noise masking, and by the limited response time that short detection range allows.

Camera-based systems

Computer vision approaches — using cameras and image processing software to detect other aircraft in the visual field — offer the potential to detect both cooperative and non-cooperative traffic using relatively lightweight and low-cost sensors. The technical challenge is significant: detecting a small aircraft at a sufficient range to allow a meaningful avoiding manoeuvre requires high-resolution imagery, robust object detection algorithms capable of operating in varying lighting and weather conditions, and processing capability that can run in real time on an aircraft platform with limited computational resources.

The field of machine learning-based visual DAA has advanced significantly over recent years, with several research programmes producing promising results in controlled conditions. Deployment in operational drone delivery systems remains limited, partly because of the certification challenges associated with machine learning-based safety systems and partly because of the reliability questions that arise in the full range of conditions a commercial operation encounters.

Standards and certification

The development of standards for DAA systems has been an active area of work by ASTM International, RTCA, and EUROCAE. ASTM F3442 covers DAA performance requirements for small UAS, and RTCA SC-228 has published minimum operational performance standards for DAA systems. These standards provide the technical basis for certification of DAA equipment and for the regulatory acceptance of DAA-equipped aircraft in authorisation processes.

The practical reality for most commercial drone delivery operators today is that DAA is addressed through a combination of operational procedures — corridor design that avoids known traffic areas, coordination with UTM systems that provide awareness of other authorised flights, and operational altitude choices that minimise conflict risk — rather than through onboard autonomous avoidance technology. Full autonomous DAA capability is the direction of travel, but operational procedure-based risk mitigation is the current reality for most operators.