From a bearing to a fix
The KrakenSDR ATAK Plugin turns the receiver into an ATAK sensor. Direction-of-arrival bearings, the power lobe and RSSI-coloured breadcrumbs draw straight on the map, every reading accumulates into a running estimate of where the emitter actually is, and the whole picture goes out as CoT. No second screen in the vehicle.
Bearings from every position accumulate into one 100-metre fix
The plugin subscribes to the KrakenSDR middleware websocket for live DOA data and falls back to polling on its own if the socket is unavailable. Location comes from this EUD's GPS, the KrakenSDR's own GPS, or a surveyed static position with the true-north bearing of element zero.
Centre frequency and VFO across 24–1766 MHz, uniform gain, antenna arrangement and element spacing, DOA processing on or off, and the algorithm — Bartlett, Capon, MEM or MUSIC. Settings are read back after applying so the operator can confirm what the device accepted.
Every reading is walked out along its azimuth in 100-metre steps to five kilometres and binned into MGRS cells weighted by received power. Cells hit from several positions win, and the leading cell publishes as a transmitter marker over CoT and multicast whenever it changes.
Direction finding is normally a two-screen job: the receiver publishes to its own web interface, the tactical picture lives in ATAK, and the operator translates between them by hand. The plugin puts the whole problem inside the map.
The DF receiver publishes to its own web interface, the tactical picture lives in ATAK, and the operator reads a bearing off one screen and estimates where it points on the other. That costs time in exactly the situation where the emitter is moving, and it keeps the fix inside one vehicle instead of on the team’s common operating picture.
The plugin subscribes to the KrakenSDR middleware websocket for live direction-of-arrival data and falls back to polling on its own if the socket is unavailable. Each reading is drawn on the map rather than in a browser: the reported azimuth, a power lobe built from the DOA spectrum, and route breadcrumbs coloured green through red by received signal strength. An averaged line smooths the last fifty angles for a steadier read on the move.
For each reading the plugin walks the azimuth outward in one-hundred-metre steps to five kilometres, bins every step into the MGRS square it lands in, and weights that square by the received power in the reported bearing bin. Squares hit from several positions accumulate score; the highest-scoring square is the current best estimate of where the emitter is, and it re-publishes as a CoT marker whenever it changes.
Everything an operator would normally open the device’s web interface for is exposed in the plugin. Centre frequency and VFO across 24–1766 MHz, uniform gain, antenna arrangement and element spacing, DOA processing on or off, and the algorithm — Bartlett, Capon, MEM or MUSIC. Settings are read back after applying so the operator can confirm what the device actually accepted, and a built-in array calculator sizes the geometry for the frequency being worked.
A DOA waterfall shows bearing against time so the operator can see how stable a track really is — it is the direction-of-arrival power spectrum, not an RF frequency spectrum. Every session is stored as a Run in an on-device SQLite database with its distance driven, time driven and final transmitter estimate, exportable to CSV and re-importable to replay the data.
Extended feature list
Subscribes to the KrakenSDR middleware websocket for live direction-of-arrival data and falls back to polling at the configured interval on its own, so it works against both the newer middleware and the original data-out server.
Take position from this EUD's GPS, from a GPS connected to the KrakenSDR, or from a surveyed static site with the manual latitude, longitude and true-north bearing of element zero entered by hand.
When the EUD is the location source, readings below a configurable speed in kilometres per hour are discarded, so a stationary vehicle does not pollute the accumulator with noise. Set it to zero to disable.
Element count, spacing, array radius and an angular-resolution estimate for the frequency being worked, with a spacing sweep and an element plot — so the geometry is sized in the plugin rather than on a spreadsheet.
Bearing 0–360° across, time down with newest at top, colour showing relative power in each bearing bin and a line marking the reported bearing. This is the direction-of-arrival power spectrum, not an RF frequency spectrum — the device does not publish frequency-domain data over the network.
Connections are managed as a list, each with its own worker thread and its own map layer, so several receivers can work the same picture from different positions.
The transmitter marker, the lobe polygon and the DOA lines are published to the connected TAK Server and to multicast, so the fix reaches every operator on the net rather than staying inside one vehicle.
Every session is stored as a Run in an on-device SQLite database with its distance driven, time driven and final transmitter estimate. Export to CSV for after-action work, or import a CSV to replay a previous run.
Mobile, static, or both at once.
The plugin does not assume a vehicle. Take position from the EUD, from a GPS on the KrakenSDR, or from a surveyed static site — and run more than one receiver against the same emitter when you have them.
Mount the array with element zero facing the direction of travel and take location from the vehicle. Set a minimum speed so stationary noise is discarded. Bearings, lobe and RSSI breadcrumbs draw as the crew drives, and the estimate tightens as readings come in from different positions.
Survey the array's position and take a true-north — not magnetic — compass reading of the bearing of element zero, keeping the EUD away from the antenna's magnetic base. Set the location source to manual and enter latitude, longitude and that bearing.
Run more than one KrakenSDR as separate connections on the same map, each with its own layer. Both vehicles' bearings and the resulting transmitter marker land on the same common operating picture through TAK Server or multicast.
Every run is stored with distance driven, time driven and the final transmitter estimate. Export to CSV for the after-action review, or re-import a CSV to walk the data again.
Stop reading bearings off a laptop.
License the plugin in under 2 minutes. Or book a scoping call if you're deploying to a team of 10+ and need a volume quote.
