Sensor types
Reach is the number people compare. What a sensor measures matters more, because it decides whether the sensor can locate anything by itself.
DomeCommand takes in whatever you have. What it does with a detection depends on what that sensor actually measured, and there are only three answers.
| Measures | Means | What fusion can do with one |
|---|---|---|
| A position | Range and bearing together | Track it on its own |
| A bearing | A direction, no range | Nothing alone. It needs a second source. |
| A zone | Something is within this area | Confirms presence. Localises nothing. |
That table is the whole sensor-selection problem. A sensor that reports a bearing is not a worse sensor, it is a different one, and a site built only from bearing-only sensors has no track at all.
The sensor types
Figures are the defaults each type ships with. Your own hardware overrides them.
| Sensor | Measures | Reach | Field of view | Notes |
|---|---|---|---|---|
| Ground radar | A position | 5 km | 360° | The workhorse. Localises alone. Struggles with small, low, slow targets and terrain clutter. |
| LIDAR | A position | 1.5 km | 360° | Very precise, short. Good as a final layer over a specific asset. |
| RF direction finder | A bearing | 8 km | 360° | Detects the control link, often before the aircraft is visible. Blind to anything not transmitting. |
| ELINT / ESM | A bearing | 12 km | 360° | The longest reach here, and no range at all. An early warning sensor. |
| Acoustic array | A bearing | 700 m | 360° | Hears rotors. Degrades badly in wind and traffic noise. |
| Acoustic node | A zone | 300 m | omni | Cheap, deployable, says only something is here. Strong in numbers. |
| EO turret | A bearing | 4 km | 12° | Identity. Where a contact stops being a radar return and becomes a known type. |
| IR turret | A bearing | 3 km | 12° | Identity at night and through smoke. |
| Long-range IR turret | A bearing | 6 km | 25° | Same job, further, wider. |
Layering a site by range and measurement
The sensors that see furthest cannot tell you where anything is. ELINT reaches 12 km and RF 8 km, and both report a direction. The sensor that localises on its own reaches 5 km.
So the shape of a workable site is usually:
- Something long-range and bearing-only for warning. You know something is out there and roughly where to look.
- Something that localises so a track can form and the planner has geometry to work with.
- Something that identifies, cued onto the track once it exists.
That is also why two cheap bearing-only sensors are often worth more than one expensive sensor. Two direction finders at different points give you an intersection, which is a position. One gives you a line.
Cueing a turret
A 12° field of view sees almost nothing until it is told where to look. That is the intended arrangement: something else detects, and the turret is cued onto the bearing. A turret left staring at a fixed sector is a turret that will miss almost everything.
Placing a sensor
A sensor that is not placed contributes nothing, and this is the most common setup mistake. A 300 m acoustic node with no position is a coverage radius with no origin. Its detections constrain nothing, no track can form from them, and its row still reads as healthy.
Aircraft are exempt. A drone knows where it is and reports it, so a camera mounted on one resolves from the aircraft's current position on every tick.
See Sites for placing them, and Adding an asset for the flow.
Testing a sensor mix in simulation
Every sensor type here exists in simulation with the same detection model. You can lay out a site, run a raid against it, and see whether the picture firms up in time to matter, before committing to hardware.
That is the intended way to choose a sensor mix. See Simulation.
Where to go next
- Integrations for the protocols a sensor can speak, and a page per protocol.
- Supported hardware for the specific products this build knows.
- Sensor fusion for what happens once several of these disagree.