Perception systems
PANOPTES is a helmet- and glasses-mounted system. It shows the soldier one quiet heads-up display and saves every sound and encounter along the way.
MTIVE Industries is a defense technology company. We build AI-powered perception systems that help the dismounted soldier take more calculated steps in a fight.
What we do
A soldier on foot can only see and hear so far, and can't remember everything that happened. MTIVE builds AI that reads camera, thermal, motion and radar data, shows the soldier only what matters, and keeps a record of every sound and encounter. The decisions stay with the soldier.
PANOPTES is a helmet- and glasses-mounted system. It shows the soldier one quiet heads-up display and saves every sound and encounter along the way.
ARGUS is MTIVE's onboard AI. It finds people in the camera, thermal, motion and radar data and decides which are worth showing the soldier.
ARGUS is trained to find people from a worn camera at eye level, the view a soldier on foot actually has.
Our main product
PANOPTES is a helmet- and glasses-mounted perception system for the dismounted soldier. It points out the people around them as they move, and saves every sound and encounter they go through, so each next step can build on what already happened.
Drag to compare · Low-light footage, ARGUS v1
Drag the handle across the scene. To its left is what the eye picks up in the dark, to its right is what reaches the soldier's glasses.
Source footage: "Airsoft inside a prison" by Names Nicco, https://www.youtube.com/watch?v=H7XcAX9JQN8. Licensed under Creative Commons Attribution (CC BY), checked 2026-10-30. Overlay (ARGUS detections, grayscale conversion) added by MTIVE. Not endorsed by the original author.
The helmet and glasses carry the sensors and the display. A compact unit on the back does all the processing. A bracelet on the forearm takes the commands.
CH 01Cameras, eye levelCH 02MotionCH 03Thermal, always onCH 04UWB radar, front-facingARGUS, MTIVE's perception AI, detects people, fuses the 4 channels and decides which returns are worth interrupting the soldier's view.
All processing on the onboard unit · No cloud · No reach-back link
Solid boxes for confirmed detections, dashed boxes for unconfirmed signals. Nothing else reaches the glass.
Gestures go to the processing unit. No voice, no buttons, hands stay on the weapon.
Every channel scans the same scene all the time. Open each one to see what it picks up, then compare it with what reaches the soldier's glasses.
Simulated scene. Illustration, not sensor footage.
View · Heads-up display
The scene through the glasses, plus only what the alert layer passes. 3 of 9 returns reach the display.
Solid box. Confirmed by more than 1 sensor.
Dashed box. Unconfirmed, a single channel only.
CH 01 · CAM
Eye-level cameras see from the soldier's own point of view. ARGUS, the onboard perception AI, finds people in the frame.
CH 02 · MOT
Catches movement too slow or too small for the human eye, such as a body moving through grass.
CH 03 · IR
Always on and scanned continuously, including while the thermal image is not on the display.
CH 04 · UWB
Front-facing, directional and aligned with the cameras. Detects presence and breathing through walls.
Bench figures for the current model. Full test conditions and accuracy results are shared in briefings.
Bench measurement on a desktop GPU.
Under the 0.04 limit MTIVE set for the model.
Nearly 7 in 10 people the eye can hardly pick out.
An EMG bracelet on the forearm reads muscle signals as gestures. There is nothing to press and nothing to say out loud.
Nothing to find or press.
Nothing to say out loud.
Gestures are read at the forearm.
Muscle signals from the forearm become commands.
Who we are
MTIVE Industries is a defense technology company. These are the people who lead it.
Every model is trained on data MTIVE owns, with no borrowed pretrained weights. Allied and sovereign customers get a stack with no foreign licensing strings and no export-control exposure.
All processing happens in the compact unit worn on the back. PANOPTES needs no cloud and no reach-back link to work.
Human-in-the-loop by design. It sharpens the operator's judgement and leaves every decision with the person wearing it.