Both validated in the official crash test report, available under NDA

3D detection and tracking output | which was the best output type of the test | we were 1 of the 4, out of 14 manufacturers that produced map coordinates, which was the highest-precision output type that was observed at the crash test. most other manufacturers managed to produce presence-only, bearing-only, or nothing at all
with a range of 65 to 140 meters | from each sensor node, with sub second detection | we had the shortest confirmed range of the whole crash test, but our sensors were also at least ~10x cheaper than the next device being tested, and ~100x cheaper than the one with the longest range which was using both acoustic and optical sensors

This is exactly what our architecture was designed for. Since we deploy a forward, dense mesh of cheap and shorter-range nodes, our effective detection range is abstracted into the length of the fiber that connects each sensor node to the hub, plus the area that the dense mesh of sensors occupies in the ground. Also, since we have a large number of overlapping sensors, we can detect faster and with superior precision, and therefore output a high quality track for downstream systems to reliably engage with those targets.

Data from the independent FPV-detection crash test, end of May 2026, Central Ukraine. Targeted at foreign and Ukrainian manufacturers fielding systems at TRL 5+ stage. Combat FPV drones were localized under live electronic-warfare jamming, rain and wind, using classical DSP (multi-sensor time-difference-of-arrival), and no neural network. Every stated figure was measured at the crash test, and is not a final specification. Scrubbed crash test data is available on request.

Snake Island Institute|0.8 3rd Army Corps|0.8 BRAVE1|0.88