Edge intelligence for strategic environments.

LENS View was built for low-bandwidth, low-power edge nodes - police booths, transit stations, remote substations. The same architecture deploys to airborne and unmanned platforms: on-device detection, classification, tracking and event triggering, with telemetry-only uplink and full degraded-mode operation when comms drop.

Tested on small-form-factor compute (Jetson Orin and equivalent), with partner-led integrations on UAS and ground-uncrewed platforms. The same models that run in a command centre run on the airframe.

Multi-modal sensing fused on the LENS substrate: EO/IR cameras, acoustic arrays and passive RF. Detect, classify and track Class 1-3 UAS with high-confidence bird discrimination. Provides the detection layer; effectors are integration-dependent and customer-procured.

Suitable for perimeter security at airports, energy sites, government complexes and event zones. Edge-native - no cloud round-trip in the detection loop.

Off-the-shelf Wi-Fi access points already light up most contested spaces. LENS Pulse reads channel state information (CSI) and infers human presence, count, posture and action class - through walls, in the dark, with no privacy-invasive optics. Built first for defence and critical-infrastructure perimeters where camera placement is impossible.

See the research write-up for the technical foundations. Strategic deployments are partner-led.

The LENS substrate produces detection, tracking and identification-confidence output suitable for use as the perception layer in human-supervised targeting workflows. Engagement decisions, rules of engagement and effector control are out of scope of LENS - they remain with the operator and the operator's sovereign command structure.

Every inference is logged with explainability surfaces and tamper-evident audit trails. We hold to the principle that consequential actions require a human decision and a paper trail.