New ArkNEW ARKACOUSTIC DETECTION

BUYER KNOWLEDGE / NEW ARK

Performance & Field Validation

Plan evidence-based evaluation of detection performance, environmental conditions, demonstrations, and written acceptance criteria.

9 questions found · of 9

35—43

Performance & Field Validation

35

What detection probability can we expect at 500 m, 1 km and 2 km?

No verified detection probabilities at 500 m, 1 km or 2 km are established here. Published range context is not a probability curve and cannot show how often a defined target will be detected at each distance. Target characteristics and site conditions must accompany any such result.

Agree the aircraft, flight conditions, distances, observation counts and detection endpoint for an authorised test. Record labelled ground truth and background activity, then report detection probability separately at each agreed condition alongside false alarms. Do not extrapolate one successful observation into a reliable probability at longer distances.

36

How does range change with drone size and payload?

Acoustic detection depends on observable motor, rotor and airframe sound, not size alone. A larger aircraft or a different payload does not establish a predictable detection-distance increase. Flight condition and the sound's propagation through the site's noise, terrain and obstructions also affect the usable observation.

Compare the specific aircraft and payload configurations relevant to your project in an authorised evaluation. Keep the distances, environmental conditions and detection definition explicit, and report results for each tested configuration. Do not convert a fixed-wing or multi-rotor catalogue reference into a universal size-to-range rule.

37

Can you detect a 150 km/h FPV drone?

A verified capability against a 150 km/h FPV drone is not established here. Acoustic sensing may observe its motor and rotor sound, but aircraft speed alone does not determine whether a usable detection or continuous track will be obtained. Noise, propagation and observation conditions remain important.

Define the exact aircraft, evaluation conditions and required outputs for an authorised field test. Measure detection, time to alert and track continuity separately with labelled ground truth. A general drone-detection description should not be interpreted as a validated high-speed target specification or an interception capability.

38

What happens when several drones approach at once?

Simultaneous drones create a multi-source assessment problem: detecting several sounds is not the same as keeping a separate, continuous track for each aircraft. The selected system's behaviour needs validation, and published concurrent-object references do not guarantee results for every group or acoustic environment.

Specify the simultaneous target count, separation and required alert and track outputs. Assess labelled observations for missed detections, false alarms and continuity, and confirm how uncertainty is communicated to operators. Any response remains separate from detection and must be reviewed by the responsible legal and safety authorities.

39

Can you detect a drone that is gliding or hovering silently?

A powered hovering rotor is not silent, although its sound may be too weak or masked to provide a usable observation at the array. A quiet gliding aircraft can be an acoustic limitation because detection depends on sound that actually reaches the sensor.

Do not treat hovering and gliding as the same test condition. Evaluate the relevant aircraft states against site noise, distance and obstruction, with detection defined clearly. Complementary radar or EO/IR can be considered where acoustic evidence is limited, rather than claiming that acoustic sensing detects every apparently silent aircraft.

40

How do we validate performance before we buy?

Use an agreed, authorised field-validation plan tailored to your facility rather than relying only on catalogue descriptions. Start with a survey of noise, obstructions, permitted installation boundaries, access and available power and networking. Define the proposed configuration and the target aircraft before testing.

Agree distances, conditions, observation counts and measurement endpoints for detection probability, false alarms, localisation error, time to alert and track continuity. Record labelled ground truth and relevant background activity. Demonstration arrangements and acceptance criteria should be confirmed through the contact page; no trial duration or pricing is established here.

41

Do you have third-party or government laboratory test results?

No specific third-party or government laboratory report is verified here. Ask whether an applicable report can be shared for the exact product and version under consideration. A laboratory name alone would not establish that its tests cover your intended targets, environment or required outputs.

Review the test scope, methods, conditions, sample counts and measurement endpoints, and distinguish laboratory results from site performance. Where the available scope does not match your application, agree additional authorised field validation. Do not substitute a product video or an unspecified endorsement for the actual test documentation.

42

How much does weather degrade detection?

No verified percentage loss or universal weather-to-range adjustment is established here. Wind, rain and changing sound propagation can affect acoustic signal quality, while terrain and obstruction interact with those conditions. Their effects should not be reduced to one constant for every target and site.

Include representative weather conditions in an authorised evaluation and record them with each observation. Compare detection probability, false alarms and localisation or tracking outputs separately. Confirm hardware operating limits as well; surviving an environmental condition and maintaining the required detection performance are different acceptance questions.

43

How do you avoid mistaking birds or machinery for drones?

AI-assisted acoustic assessment can be part of evaluating sound sources, and multi-sensor corroboration can help assess uncertain observations. Neither establishes a guarantee that birds or machinery will never produce a false alert. Detection and classification should be evaluated as separate outputs.

Include normal machinery activity and observed bird activity in labelled validation data, alongside target-drone observations. Define what counts as a false alarm and measure it over an agreed period. Ask how confidence and uncertain classifications are presented to operators rather than assuming every alert confirms a drone's identity.

NEW ARK / PROJECT CONSULTATION

Have a question about your site?

Share your application, target types, existing sensors and procurement requirements. We can use those details to scope the next project discussion.

Discuss your project