Low-altitude security programmes usually start with one sensor and a range claim, and that is where most of them stall. A spectrum detector that sees every DJI telemetry burst goes quiet when a waypoint-flown airframe crosses the fence with its radio off. A radar that tracks everything above the treeline reports birds and rain until the operator stops trusting the alarm. And EO/IR identifies, but only if something else already told it where to look.
The answer is a layered detection architecture: passive RF for the emitter, radar for the airframe, EO/IR for identification, and one track manager fusing all three. This guide sizes that stack against published specifications from production counter-UAS hardware.
Why a single sensor always loses
RF detection is cheap, passive and information-rich: it reads the control link, the video downlink, the operator position and often the protocol fingerprint. Its blind spot is structural — a pre-programmed airframe or a fibre-optic FPV radiates nothing. Radar is emission-agnostic and gives range and velocity, but small composite airframes at 2–5 km sit close to clutter and always need confirmation. EO/IR is the only layer producing evidence a duty officer will accept, yet its field of view is measured in degrees.
What each layer contributes in practice
Passive RF spectrum layer
This layer answers “is there a drone radio in my airspace, what band, and where is it coming from?” Wide instantaneous bandwidth matters more than raw sensitivity: a 20 MHz–8 GHz front end with direction finding across 300 MHz–8 GHz covers hobby links, FPV video, cellular-based control and 5.8 GHz digital FPV without re-tuning. Airframe classification from the RF fingerprint, plus Remote ID and ADS-B, gives a first identification before any radar return exists.
Radar layer
Radar contributes what RF cannot: a track on a non-emitting target, with azimuth, elevation, range and closing velocity. Azimuth accuracy of about 0.4° and elevation of about 0.5° with 10 m range accuracy is enough to hand a cue to an EO/IR payload at a few kilometres.
EO/IR confirmation layer
Visible-light identification to roughly 3 km and infrared to roughly 1.5 km is the practical envelope for a fixed mast — enough to read airframe type, payload and whether a person is underneath. Without this layer, every engagement decision rests on a spectrum alarm.
Reference specification set
These figures come from production units and show the spread between a detection-only node and an integrated site system.
- URD360 Pro spectrum detector: detection 20 MHz–8 GHz, direction finding 300 MHz–8 GHz, ≥10 km range (environment and airframe dependent), ≤3° RMS DF accuracy, 360° azimuth, false alarm ≤1%, ≥30 simultaneous drones, FPV video interception plus RID and ADS-B, ≤35 kg, IP66.
- URD320 fixed detector: 845 MHz, 915 MHz, 1.2/1.4/2.4/3.3/5.2/5.8 GHz bands with 300 MHz–6 GHz expansion, ≥2 km on DJI telemetry, <3 s response, ≥10 drones, ≤40 W, ≤5 kg, IP66 — the unattended 24/7 node.
- URS680 integrated site system: 5 km detect-and-jam, EO tracking 3 km visible / 1.5 km IR, 300 MHz–6 GHz detection and jamming, GPS L1, BDS B1, GLONASS R1 and Galileo E1 spoofing, radar azimuth ≤0.4°, elevation ≤0.5°, range 10 m, ≤600 kg.
- URJ240 planar array: 845/915 MHz, 1.2/1.4/1.5/2.4/3.3/5.2/5.8 GHz, ≥3 km against a Mavic 3-class airframe at 2:1 J/S, 1 s response, 360°, IP65, directional or omnidirectional mode.
Matching the stack to the site class
High-value fixed facility
Energy sites, nuclear plants and regulatory centres justify an integrated system: one unit detecting at 5 km, identifying optically and denying the link, plus distributed perimeter detectors. The trade-off is mass and mains power.
Mobile or temporary protection
Vehicle and tripod deployments optimise for mass and power. A 4 kg node at 2 km detection is realistic for a convoy or event perimeter; it will not hold a 10 km line and should not be specified as such.
FPV-specific threat
Against fibre-optic or pre-programmed FPV the RF layer is degraded by design. Budget radar plus EO cueing, and a jammer whose response time is measured in seconds.
Commissioning details that decide real outcomes
Three details separate a working site from a demo: a shared time base so tracks from different sensors correlate, a maintained RF signature database with allowlists for your own fleet, and a written engagement authorisation for jamming and spoofing. Because spoofing GPS L1, BDS B1, GLONASS R1 and Galileo E1 together disturbs survey and timing equipment on site, plan exclusion zones before the first test.
FAQ
Can RF detection alone secure a site?
No. It is blind to autonomous waypoint flight and to fibre-optic FPV. Use it as the wide-area tripwire, not the sole sensor.
What detection range should I put in the requirement?
State the conditions: airframe class, line of sight, and whether the figure is spectrum-based or protocol-based. A 10 km DF figure and a 2 km telemetry figure can both be honest for the same product.
Do I need radar if I already have 8 GHz direction finding?
Yes, for non-emitting targets and continuous tracking through the terminal phase. DF gives bearing; radar gives a track.
Next step
Send your site layout, expected airframe classes, and power and mast constraints; we will return a sensor mix with the coverage each layer actually delivers, sized as one system rather than three quotes.
Related hardware: URD360 Pro 20 MHz–8 GHz drone detection equipment, URS680 5 km integrated defense system, and URJ240 countermeasure planar array.
Related reading
Teams that pair detection with a hard-kill layer should also review the 5 km integrated counter-UAS defense system with radar tracking and navigation spoofing specification, which shows how detection, EO tracking and jamming are packaged in one mast-mounted unit.