Assured Spectrum
Release: Beta — suitable for trial and evaluation; APIs may change before GA
Training-free RF pulse deinterleaving for contested EW/SIGINT
Assured Spectrum separates interleaved pulse descriptor word (PDW) streams in contested-spectrum environments using PRI-hypothesis coherence and bounded per-pulse influence — the same bounded-influence family as Assured Signal.
This is the deinterleave step upstream of emitter ID — not modulation recognition, not a neural classifier. Input: interleaved PDW arrays (TOA, RF, PW, amplitude). Output: per-stream labels and diagnostics. No training required.
Validation shows strong separation on dense contested PDW scenarios where legacy deinterleavers fail.
What it does
Electronic warfare and SIGINT receivers output pulse descriptor words (PDWs) — time of arrival, RF, pulse width, amplitude — from many emitters interleaved in one stream. Before you can match pulses to a threat library, you must deinterleave: assign each pulse to an emitter stream.
Assured Spectrum is that deinterleave step. It is training-free: when a new emitter appears, you do not retrain a classifier. It uses PRI-hypothesis coherence with bounded per-pulse influence (same family as Assured Signal). It is not emitter ID and not modulation recognition.
When to use it
Install in the ES/RWR/SIGINT front-end between pulse detection and threat-library correlation.
See Examples for runnable code.
Who it is for
- EW/SIGINT engineers processing interleaved PDW streams from ES/RWR front-ends
- Programs needing training-free separation when unknown emitters appear
- Embedded tactical systems requiring on-premise pulse stream separation
What ships
- Python wheel with Ed25519 licensing (
assuredspectrum) - C library (
libassuredspectrum) withdeinterleave()API - Guided journeys 01→06 under
python/examples/journeys/