Frequently asked questions

A sound field is described by two quantities. Acoustic pressure is a scalar — a single number telling you the pressure difference at a point. Acoustic particle velocity is a vector — it tells you the direction in which air particles at that point are moving. For comparison: if pressure is the acoustic equivalent of voltage, particle velocity is the acoustic equivalent of current. Classic microphones measure only pressure. Direct measurement of acoustic particle velocity in air has been possible only since the 1990s.

An acoustic vector sensor is a four-channel probe: one pressure transducer and three orthogonally placed acoustic particle velocity sensors. The pressure transducer determines what kind of acoustic event is happening. The particle velocity vector points to where the event comes from — bearing and elevation — in a single point in space.

Yes, in essence. A vector sensor captures sound in a full spherical bubble around itself. Conventional directional microphones have a narrow field of view; pressure arrays have a panoramic view but require spacing. A vector sensor has the best possible field of view in this respect.

A pressure array needs at least two microphones at some spacing — direction is derived from time-of-arrival delay. The spacing determines the optimal frequency. A high-frequency shot requires a compact array of around half a metre. A low-frequency blast requires spacing of tens of metres. A vector sensor, by contrast, captures phase and amplitude information in a single point and covers the entire audio bandwidth.

Pressure detection systems are by design dedicated to a certain kind of event — their geometry is optimised for a specific frequency band. A vector sensor is broadband: it detects low and high frequencies in the same node. This allows a single sensor type to cover the spectrum from blasts to shots and tonal sources such as vehicles or helicopters.

Vector sensors have low weight, small size and low power consumption. This allows deployment on vehicles, stand-alone ground systems, small airborne platforms and dismounted soldiers. Low SWaP is the basic precondition for a wide range of deployments.

Pressure systems have an optimal signal-to-noise ratio only in a very small part of the spectrum. Most real acoustic events have a signature covering a wider part of the spectrum. A vector sensor is broadband, so presence and direction of an event can be confirmed across multiple parts of the spectrum at once. This lowers the false alarm rate.

The lower the frequency, the larger the spacing between transducers in a pressure array must be. In practice, distributed systems are used where all data is collected in a central unit — a collaborative mode. A vector sensor captures complete directional information in a single node, so it can operate in a non-collaborative mode. This brings higher processing robustness, lower data transfer requirements and possibilities for covert operation.

Pressure transducers in a horizontal plane give bearing information. In several layers they also give elevation, but they cannot distinguish direct signals from ground reflections. A vector sensor captures in a single point all signals that travel the shortest path between source and sensor. This is important for accurate direction in a complex environment.

In pressure systems, signals from all sources and reflections are summed in each microphone before being compared with data from other microphones. This causes inaccuracies. A vector sensor captures in a single point all signals that travel the shortest path between source and sensor — reducing inaccuracies with multiple simultaneous sources.

It depends on the scenario and on the sophistication of the processing algorithms. Relevant are the loudness of the source, terrain conditions and meteorology — mostly wind direction and force. As a rule of thumb: what a human can hear in infancy, a vector sensor can detect. Specific range figures are always scenario-dependent and should not be presented as fixed values.

Yes. They do not illuminate a target and cannot be detected electronically. They do not disclose their presence. Passive sensors consume far less power than active sensors.

Yes, they cannot be jammed electronically. They are also resistant to spoofing — emitting electronic signals to confuse the opponent. Compared with radar, this is a distinctive benefit, because radar systems are jammeable and have been deliberately blinded in the past.

They meet the criteria of small size, low weight and low power. Airborne acoustic detection has the additional advantage that there are no obstacles or reflections between source and receiver. Vector sensors can also be used for two other functions crucial to UAVs — Hear & Avoid and automated take-off and landing.

DroneEAR AI is a fusion layer. It processes acoustic events together with RF detections, telemetry, ADS-B and network Remote ID. Each source contributes normalised detection events — source, coarse zone, confidence value, bearing, time. The fusion layer correlates these events and produces a threat assessment only when independent sources agree.

No. DroneEAR AI is a working concept demonstrator. Drone-threat signals are simulated. Live public feeds are ADS-B and network Remote ID, which shows only drones that identify themselves, as context. The system does not detect non-cooperative drones and is not an operational system. It is intended for technical discussion and partnership.

The false alarm rate is measured in real deployment at a specific site with specific sensors. DroneEAR AI currently has a persistent record of all assessments and alerts, after-action replay of any time window, and operator adjudication of every alert as real, false or unknown. A report from these verdicts forms the false alarm rate. No rate is claimed yet — the measurement tool is ready, but measurement requires a pilot.