EW-Sentinel

EW-Sentinel is an intelligent electronic warfare simulation system written in Rust. It models a closed-loop EW workflow: generate radar emitter signals, receive and process SIGINT captures, classify threats, generate countermeasures, package them, and deploy updates to a friendly aircraft fleet.

The project is a simulation. The CNN classifier and AI waveform generator are initialized in-process for demonstration and workflow modeling; they do not load trained operational models.

Simulated Scenario

The default scenario is Contested Airspace - IADS Penetration:

• A friendly package penetrates an integrated air defense system.
• Enemy emitters include SAM, early-warning radar, fire-control radar, AAA radar, and unknown/novel emitters.
• Friendly aircraft include platforms such as F-35, EA-18G, F-15EX, and UAVs.
• Each simulation cycle detects emitter pulses, classifies threats, generates jamming waveforms, and pushes EW package updates to aircraft.

Additional scenarios live in scenarios/, for example coastal_defense.json.

Architecture

flowchart TD
    A[Scenario JSON<br/>emitters + friendly aircraft + area] --> B[SignalEnvironment]
    C[Threat Library JSON<br/>radar/SAM signatures + recommended CM] --> B
    C --> H[ThreatClassifier]

    B --> D[Emitter Simulation<br/>pulse/IQ generation]
    D --> E[Propagation<br/>path loss + AWGN]
    E --> F[Receiver]
    F --> F1[Digitizer<br/>receiver noise + ADC quantization]
    F1 --> F2[PDW Extractor<br/>freq/bandwidth/PW/AOA/modulation]
    F1 --> F3[FFT/STFT Spectrogram]

    F2 --> H
    F3 --> H
    H --> H1[CNN Classifier]
    H --> H2[Library Parametric Match]
    H1 --> I[Threat Tracks<br/>known or novel + threat level]
    H2 --> I

    I --> J[Technique Selector]
    J --> K[Countermeasure Generation]
    K --> K1[Noise Jamming<br/>barrage / spot]
    K --> K2[Deceptive Jamming<br/>RGPO / VGPO]
    K --> K3[DRFM<br/>repeat / modulated]
    K --> K4[AI Waveform Generator<br/>fallback]

    K --> L[Package Builder<br/>version + manifest + checksums]
    L --> M[OTA Manager]
    M --> N[Friendly Aircraft Fleet<br/>EW suite updated / failed / rollback history]

    N --> O[EwState]
    I --> O
    L --> O
    O --> P[TUI Dashboard / Headless Output]

Loading

Runtime Loop

sequenceDiagram
    participant Main
    participant Loop as EW Loop
    participant Env as Signal Env
    participant Rx as Receiver
    participant AI as Threat Analysis
    participant CM as Countermeasure
    participant Deploy as Deployer
    participant UI as Dashboard

    Main->>Loop: load scenario + threat library
    loop each cycle
        Loop->>Env: tick simulated time
        Env-->>Loop: IQ captures from emitters
        Loop->>Rx: digitize + extract PDW + spectrogram
        Rx-->>Loop: PDWs + spectrograms
        Loop->>AI: CNN + library fusion
        AI-->>Loop: threat tracks + novel count
        Loop->>CM: select techniques + generate waveforms
        CM-->>Loop: jamming waveforms
        Loop->>Deploy: build package + push OTA
        Deploy-->>Loop: deployment records
        Loop->>UI: stage/threat/waveform/deploy events
    end

Loading

Module Map

• src/main.rs: CLI entrypoint, scenario loading, threat library loading, dashboard startup.
• src/config.rs: CLI arguments and scenario JSON schema.
• src/pipeline/loop_runner.rs: main EW simulation loop.
• src/signal_env/: emitter creation, modulation generation, propagation effects.
• src/receiver/: digitization, PDW extraction, FFT spectrogram generation.
• src/threat_analysis/: CNN inference, threat library matching, unknown threat characterization.
• src/countermeasure/: technique selection and jamming waveform generation.
• src/deployer/: deployment package building, OTA update simulation, rollback helpers.
• src/models/: shared signal, threat, aircraft, countermeasure, and deployment types.
• src/ui/: terminal dashboard and headless output.
• src/ml/: lightweight neural network layers and tensor utilities used by the classifier/generator.

Requirements

• Rust toolchain with Cargo.

The project uses dependencies from Cargo.toml, including tokio, serde, ratatui, crossterm, ndarray, and rustfft.

Running

Run from the project root:

cargo run

Run without the TUI dashboard:

cargo run -- --no-dashboard

Choose a scenario:

cargo run -- --scenario scenarios/coastal_defense.json

Set cycle count, seed, and time compression:

cargo run -- --max-cycles 5 --seed 123 --time-compression 60 --no-dashboard

Build release binary:

cargo build --release

CLI Options

--scenario <PATH>          Scenario JSON path
--max-cycles <N>           Override scenario cycle count
--seed <N>                 RNG seed for reproducible runs
--time-compression <N>     Simulation time compression factor
--no-dashboard             Print headless progress instead of TUI
--debug                    Enable debug logging
--threat-library <PATH>    Threat library JSON path

Data Files

data/threat_library.json contains known threat signatures:

• frequency range
• typical PRI and pulse width
• modulation type
• scan type
• lethal range
• recommended countermeasure techniques
• classifier label

Scenario files in scenarios/ define:

• operating area
• emitter groups and modes
• friendly aircraft
• simulation cycle count
• simulated duration per cycle

Notes

• Emitter positions and signal parameters are randomized from scenario and threat library inputs.
• Unknown emitter types are modeled as novel threats and characterized heuristically.
• OTA deployment includes simulated failures such as degraded data links and version incompatibility.
• The dashboard is rendered with ratatui; press q to quit.