QQYLAIDiscuss an integration ↗
RF-NATIVE TRACK FUSION / COUNTER-UAS

Many sensors.
One trustworthy
picture of the sky.

Qylai associates passive-RF bearings with radar and EO/IR detections into one coherent track picture—so counter-drone systems can follow dense activity without duplicate, conflicting, or unexplained tracks.

Every result is auditable, replayable, and ready for the ATAK and C2 systems customers already use.

See how Qylai works ↓Talk with the team ↗
RF-NATIVE / AUDITABLE / C2-COMPATIBLE
WHAT QYLAI DOESSAME AIRSPACE / THREE SENSORS
01CONFLICTING REPORTS
RADAR
3 returns
PASSIVE RF
bearing cue
EO / IR
visual confirm

One object. Three incomplete views.

02QYLAI
Associate the same objectWeight sensor uncertaintyPreserve the evidence
DETERMINISTIC FUSION LAYER
03ONE TRUSTED TRACK
TRK-017IDENTITY STABLE
CONTRIBUTORSRADAR · RF · EO
EVIDENCEATTACHED

Ready for existing C2 and evaluation tools.

EXISTING SENSORS INQYLAI DOES NOT REPLACE THE SENSOR OR C2TRUSTED TRACKS OUT
THE COUNTER-UAS PROBLEM

More sensors do not automatically create more truth.

Counter-drone teams increasingly combine radar, RF, EO/IR, and cooperative feeds. The hard part is deciding which observations belong to the same object—and proving the resulting track is trustworthy.

01

Sensors disagree.

Radar, passive RF, and cameras observe the same airspace differently. Their detections arrive with different timing, uncertainty, and blind spots.

02

Density breaks identity.

Crossings, dropouts, and coordinated motion create duplicate tracks and identity swaps—exactly when the operator needs a stable picture.

03

Results are hard to defend.

When a fused result cannot be replayed or traced to its source observations, buyers cannot compare configurations or explain failures.

BUILT FORSensor OEMs

Show how a radar, passive-RF, or EO/IR product improves a fused track picture—without building fusion in-house.

BUILT FORATAK teams, integrators, and evaluators

Publish scored, provenance-carrying tracks into existing C2, and reproduce every test from the detection log.

QYLAI IS NOTAnother sensor or C2 suite

It is the RF-native fusion and evidence layer between the sensors and the systems already responsible for decisions and response.

THE PRODUCT

Conflicting detections become one defensible track.

Qylai resolves which observations belong together, combines their uncertainty, and publishes a stable track with the supporting evidence attached.

01

Observe

Existing sensors report partial views of the same airspace, each with its own timing and uncertainty.

KEEP THE SOURCE EVIDENCE
02

Associate + fuse

Qylai identifies observations of the same object and weights each contribution by what that sensor actually knows.

LEGIBLE ESTIMATION
03

Publish

Existing C2 and evaluation tools receive one coherent track instead of competing sensor reports.

OPEN OUTPUTS
WHY QYLAI

Trust is not a dashboard feature.
It is the product.

01INSPECT IT
TRK-017 / EVIDENCE3 SOURCES
RADAR-0414:02:31.2040.42
RF-1214:02:31.2170.31
EO-0314:02:31.2380.27

See why a track exists.

Every fused track points back to the sensor observations that shaped it, so an integrator or evaluator can inspect the evidence instead of trusting a black box.

EVIDENCE LINEAGE
02REPLAY IT
SAME INPUT / SAME RESULTMATCH
RUN A7E3A—91C4
=
RUN B7E3A—91C4

Reproduce every test.

The same recorded scan sequence recreates the same result. Teams can compare releases, investigate failures, and prove what changed.

DETERMINISTIC REPLAY
03INTEGRATE IT
ONE TRACK CONTRACTOPEN OUTPUTS
TRK-017state · uncertainty · evidence
CoTGEOJSONSAPIENT

Keep the existing stack.

Qylai sits between sensors and command systems. Open outputs let customers improve the track picture without replacing maps, workflows, or response tools.

OPEN INTERFACES
KNOWN-TRUTH BENCHMARK

Three sensors beat the best one alone.

In this known-truth scenario, Qylai produced 47% less aggregate tracking error than the strongest individual sensor. That means a better balance of correct locations, fewer misses, and fewer false tracks.

WHAT GOSPA MEASURES

GOSPA is a standard multi-object tracking score. Lower is better: a system is penalized for wrong locations, missed objects, and false tracks—not just the final object count.

MEAN AGGREGATE TRACKING ERROR / LOWER IS BETTERKNOWN-TRUTH RUN 001
BEST INDIVIDUAL SENSORbaseline
16.8
QYLAI FUSIONradar + RF + EO/IR
8.8
↓ 47% AGGREGATE ERRORREPRODUCIBLE FROM DETECTION LOG
THE COMPLETED PLATFORM

RF-native fusion.
Defensible evidence. Open edges.

One product surface for sensor OEMs, ATAK operators, and evaluators: associate heterogeneous detections, prove the result, and publish into the stack already in the field.

01FUSION CORE

RF-native association.

Associate passive-RF bearings with radar and EO/IR detections into one coherent track—identity preserved through crossings, clutter, and dropouts.

  • Bearing + position fusion
  • Legible deterministic estimation
  • Track identity at density
02EVIDENCE LAYER

Every track can be defended.

Provenance on every estimate, scan-exact replay from the detection log, and scored evaluation so OEMs and T&E teams can prove what the fused picture actually did.

  • Per-track evidence lineage
  • Bit-for-bit deterministic replay
  • GOSPA-scored known-truth runs
03OPEN EDGES

Ship into the stack you already run.

Publish into ATAK/CoT and open architectures. Sensor OEMs and integrators keep their hardware and C2; Qylai improves the track picture underneath.

  • CoT / ATAK outputs
  • OEM sensor adapters
  • SAPIENT-compatible interfaces
SENSOR OEMS / ATAK TEAMS / EVALUATORS

Bring a sensor stack or recorded dataset.
See the fused ATAK picture.

Qylai works with counter-UAS OEMs and integrators who need RF-native association, scored replay, and tracks that land in the C2 they already run.

DISCUSS AN INTEGRATION↗
UNCLASSIFIED DISCUSSIONS ONLYQYLAI.COM / 2026