Big Auto One / Sensing
See the threat before it arrives
Big Sensor One is our sensor-fusion software. It brings radar, RF detectors, cameras and ground sensors onto one map, fuses what they report into one track per object, finds and ranges drones, and points a camera at them. It runs on a single box with no internet and no cloud, and it is designed to ride on the Spectre as a mobile sensor node.
Drone tracking
Every drone,
one track
Small drones show up differently to every sensor: a return on radar, a control link on an RF detector, a broadcast identity, a speck on a camera. Big Sensor One puts those reports together so the operator sees one drone, not four alarms.
Multi-sensor drone tracks
Radar, RF detection, broadcast identity and cameras are fused into a single track per drone, with duplicate and ghost tracks removed.
Demonstrated in simulation
Who is flying it
Where a drone broadcasts or transmits its identity, the track shows make, model and serial number and the operator's reported position on the map. A broadcast alone never marks a drone as friendly.
Demonstrated in simulation
Bearing-only tracking
Direction-finding RF detectors that report only a bearing still produce a track when two or more of them hear the same drone.
Demonstrated in simulation
Silent drones
A drone that transmits nothing is still tracked by radar and cameras, and is flagged as having no identity rather than ignored.
Demonstrated in simulation
Altitude and height
Every air track carries its altitude and its height above the ground beneath it, worked out from offline terrain.
Demonstrated in simulation
Through cover
When a tracked object passes behind a building or tree line, its track is held and predicted instead of dropped, so it comes back as the same track.
Demonstrated in simulation
Threat ranking
Inbound drones are ranked by urgency, with a countdown to arrival, so the most pressing threat is always at the top of the list.
Demonstrated in simulation
Cooperative traffic
Aircraft and ships that report their own position are shown alongside everything else, so known traffic is not mistaken for a threat.
Demonstrated in simulation
Ranging and cueing
Where it is,
and a camera on it
Knowing a drone exists is the start. The crew needs its range and direction from where they are standing, and eyes on it to decide what it is.
Range, bearing, elevation
Every track carries its range, bearing, elevation and closure from the site, and from any user's own position.
Demonstrated in simulation
Range rings
Rings around the site on the map, and an identification ring inside which an unidentified drone is marked as a suspect.
Demonstrated in simulation
Cue card on a phone
A handheld card shows the direction, elevation and range to a chosen track, or to the most urgent threat, from the user's own position, with no network needed on the phone.
Demonstrated in simulation
Slew-to-cue
When one sensor detects something, a pan-tilt-zoom or day/thermal camera is pointed at it automatically. If the camera does not find it, the camera searches around the cue.
Demonstrated in simulation
Click a drone, see it
Select a track and the gimbal camera turns and zooms onto it, with its direction and range on screen.
Demonstrated in simulation
Position quality, shown
Each position is shown with its uncertainty, and higher-quality coordinates are visually distinguished from rough ones.
Demonstrated in simulation
Ground and perimeter
People and vehicles,
as well as drones
The same picture covers what moves on the ground around a site, a route or a halted vehicle.
Camera analytics
Live video from fixed and pan-tilt-zoom cameras, with people, vehicles, aircraft and boats detected and classified on the stream. Government or third-party detection models can be dropped in.
In development · first real camera connected
Unattended ground sensors
Motion and seismic ground sensors, and other sensor types, report onto the same map and into the same tracks.
Demonstrated in simulation
Zones drawn in 3D
Operators draw alert and exclusion zones on the map, with a floor and a ceiling, to focus alerts and suppress false alarms.
Demonstrated in simulation
Detection and threat alerts
Rules turn detections, classifications and zone crossings into alerts, each with a one-touch replay of the video and map at that moment.
Demonstrated in simulation
Weather-aware
A weather overlay on the map, taken into account when cueing cameras.
Demonstrated in simulation
One picture
Map and video,
in step
The operator works from one display: the fused tracks on a map and the video that shows them, kept in step live and in replay.
Fused map and video
Tracks in standard military symbols on the map, with track trails drawn on both the map and the video, and single, side-by-side, picture-in-picture or quad video.
Demonstrated in simulation
Offline maps and 3D
Vector maps, satellite imagery, elevation and a 3D view, all held on the box. Nothing is fetched from the internet.
Demonstrated in simulation
Synchronized timeline
Replay all video and sensor data together on one timeline, and export a time range as a clip package that plays on any computer without installing anything.
Demonstrated in simulation
Many nodes, one view
Several Big Sensor One nodes report to one view, each with its own status and history.
Demonstrated in simulation
Runs on one box, disconnected
Everything runs on a single edge processor with no cloud, internet or license server. Workstations and phones are only displays.
In development
Open standards
Into the picture
you already run
Big Sensor One is built on open, government and public standards, so sensors plug in and tracks flow out without a proprietary middle layer.
TAK and GeoJSON
Fused tracks published to TAK and as GeoJSON, with standard metadata on recorded video.
In development
SAPIENT sensor network
Accepts third-party sensors that speak the open SAPIENT standard and fuses them onto the same tracks, reports up to a higher fusion node, and can run any of its own sensors as a standalone node.
Demonstrated in simulation
A graph of everything
Every detection, track, event and relationship is stored in a graph database with a documented API for analysis and third-party tools.
Demonstrated in simulation
Sign-in and audit
CAC/PIV and single sign-on, permissions by role, account lockout, and a tamper-evident audit log of every action.
In development
On the Spectre
A sensor node
that can move
Big Auto One already takes hazards from the common operating picture and turns them into keep-outs the vehicle routes around. Big Sensor One is designed to close that loop on the vehicle itself.
Spectre as a sensor carrier
Big Sensor One running on the Spectre, so a perimeter can be moved, extended or put where it is needed and taken away again.
Planned
Tracks become keep-outs
Threats seen by Big Sensor One reach the vehicle as keep-outs in its authorization, so it re-routes before it reaches them.
Planned
Companion drones as sensors
The Spectre's own small air and ground drones feeding their sightings into the same fused picture.
Planned
Your site
Show us what you need to see
Tell us the site, the route or the threat. We will build it in simulation on the real ground and show you what Big Sensor One sees.