The Low-Slow-Small Detection Problem
Every detection principle fails somewhere. Knowing exactly where is the beginning of system design.
Why LSS targets are hard
A 35 cm multirotor flying at 100 meters is close to the hardest object in modern surveillance: radar cross-section around 0.01 m², speed from zero (hovering) to 100 km/h, altitude below the treeline. Every classical sensor was built for something else — aircraft, emitters, vehicles — and each fails on a different axis.
The four principles, with their failure modes
Radar — sees everything, says nothing
Excellent coverage and all-weather operation, but small RCS drowns in ground clutter, hovering targets vanish into Doppler filter notches, and multipath plagues urban sites. Worse, a transmitting radar is itself visible and jammable. And a radar plot is a point, not an identity.
RF spectrum — hears only speakers
Detects control and video links at kilometers, direction-finds the aircraft and localizes the pilot. But its premise is emission: a radio-silent, pre-programmed or home-built aircraft is invisible to it, complex RF environments raise false-alarm rates, and multipath reflection degrades localization.
Acoustic — cheap, and mostly theoretical
Microphone arrays need no emission and work in radio silence, but background noise limits them severely; the principle remains largely experimental for operational perimeters.
EO/IR — sees and identifies, passively
Thermal plus visible imaging detects any aircraft in line of sight, day and night, with zero emissions — meaning it sees radio-silent targets, does not disturb the electromagnetic environment and preserves signatures (imagery) as evidence. Its limits: weather attenuation and the need for a cue to search efficiently.
Why fusion wins
The failure modes above are complementary: what radar misses in identification, EO supplies; what EO misses in wide-area coverage, radar supplies; what both miss in emission intelligence, RF supplies. A fused architecture runs radar as the wide-area cue, RF as the identification-and-pilot layer, and EO/IR as the positive-ID and evidence layer — each covering the others' blind spots, correlated in one command picture.
This is why our doctrine is detection-first and passive-leaning: the passive sensors (EO/IR, RF) carry the identification burden and work in the most constrained environments, while the active layer (radar) supplies the volume. Neutralization choices sit on top of this stack and are made by humans, under law.
Designing a detection layer?
Terrain, clutter and threat profile determine the sensor mix. Send yours.
Ask technically