simulation

Project ARGUS — Godot Intercept Simulation

Phase 1 of the ARGUS deliverable per Project Description.pdf: a custom Godot-based simulation used to train and validate the sensor-fusion and interception algorithms.

Run it

1. Install Godot 4.3+ (Standard, not the .NET build).
2. Godot → Import → select simulation/project.godot.
3. Press F5 (or the Play button). Main scene is scenes/main.tscn.

What's modeled

PDF requirement	Implementation
Stratospheric solar glider, 60–90 kft loiter, multi-day endurance	argus_drone.gd — constant-rate loiter at mid-band altitude; cruise speed = 28 m/s
Hypersonic cruise missile target	hcm_target.gd — Mach ~5 cruise with vertical band 18–25 km
Adversary evasion heuristics: randomized glide maneuvers, erratic altitude changes	_step_evasion() re-rolls lateral g-bias and altitude target on a randomized cadence
RF blackout from plasma sheathing	plasma_blackout flag engaged on speed × maneuver-load; HUD warns; fusion adds an IR+acoustic confirmation bonus during blackout
Multi-modal sensors (IR, EO camera, air pressure / acoustic)	sensor_swir.gd, sensor_eo.gd, sensor_infrasound.gd
Tiered detection / power-gated fusion: infrasound → SWIR → satellite uplink only on confirmed threats	power_manager.gd — IDLE → DETECT → TRACK → ENGAGE state machine with hysteresis; draw_w shown live
Sensor fusion AI estimating target state without exact coordinates	sensor_fusion.gd — inverse-variance circular bearing fusion + range-from-modality + EMA tracker. Reads only sensor outputs, never the truth state
Cost feasibility from COTS sensors / passive platform	Reflected in modeled per-sensor power budget; ENGAGE-only uplink
Clean cool UI	hud.gd — code-drawn mission HUD with tacmap, sensor stack, tier banner, blackout warning

Controls

• R — respawn HCM with new random heading / altitude profile
• T — toggle the truth marker (red sphere) for ground-truth comparison
• C — cycle camera (orbit / chase ARGUS / chase HCM)

Files

simulation/
├── project.godot
├── icon.svg
├── scenes/main.tscn
└── scripts/
    ├── sim_constants.gd     # world units, alt bands, tier thresholds
    ├── sim_controller.gd    # orchestrates sensors, fusion, power each tick
    ├── argus_drone.gd       # passive solar-glider loiter
    ├── hcm_target.gd        # hypersonic adversary + evasion + plasma model
    ├── sensor_infrasound.gd # always-on tripwire (long range, low SNR)
    ├── sensor_swir.gd       # IR thermal-verification (mid power)
    ├── sensor_eo.gd         # EO camera (high fidelity, short range)
    ├── sensor_fusion.gd     # multi-modal estimator
    ├── power_manager.gd     # tiered power-gated state machine
    ├── orbit_camera.gd      # cinematic camera
    └── hud.gd               # mission-control overlay

The estimator surface is the seam where the Phase-3 trained policy will plug in: today it's an analytic inverse-variance fuser; the same update() signature can be backed by a neural model trained on simulation rollouts.