CASPER 4 — Deterministic Governance-Only Autonomy Console Synthetic ISR telemetry · Human-gated decisions · Non-kinetic autonomy primitives CASPER 4 is a deterministic, governance-only autonomy simulator for research into safe supervisory control, risk/clarity reasoning, and human-in-the-loop decision frameworks. All telemetry is synthetic. No weapon control, no targeting, and no kinetic behaviors exist anywhere in the system.

✨ What CASPER 4 Is A self-contained simulation environment that models: • High-speed flight telemetry (fully synthetic, deterministic) • Area-of-operations (AO) geometry and threat evolution • Swarm motion within safe, bounded envelopes • Clarity / risk engines for ISR governance research • Human-gated proposal review and audit logging • CRS-1, a non-kinetic autonomy primitive that reshapes AO corridors safely and reversibly CASPER 4 is intended for research, HCI experiments, and governance prototyping, not flight systems.

🔒 Safety & Scope CASPER is intentionally constrained: • No fire control or weapon domain • No command of real systems • No targeting logic • No autonomy that operates actuators • Only geometry, telemetry, classification, and governance Every operator action is gated, audited, hashed, and fully reversible.

🧩 Architecture Overview casper4/ │ ├── app.py # Unified Streamlit console ├── CRS-1 # Corridor Re-Shaper autonomy primitive ├── Deterministic RNG # Per-domain keyed sequences ├── Telemetry Engine # Pure synthetic high-speed flight model ├── AO Model # Boundaries, threat, civ density, drift ├── Swarm Model # Non-kinetic, bounded movement ├── Risk/Clarity # Interpretability-oriented classifiers └── Governance Layer # Proposals, gating, audit chain (SHA-256) The engine is deterministic per (rng_seed, config, tick) and supports future replay/divergence testing.

🧠 CRS-1 — CASPER’s First Autonomy Primitive CRS-1 (Corridor Re-Shaper) adjusts the AO geometry based on approved proposals: • CONTRACT — tighten AO for civilian protection / high threat • EXPAND — increase AO when clarity is high and threat is low • SHIFT — small bounded nudges when nav drift warrants it CRS-1 never affects flight control. It only updates geometric governance constraints.

📦 Installation CASPER 4 requires Python 3.10+. pip install -r requirements.txt Dependencies include: • streamlit • numpy • pandas • pydeck • matplotlib

🚀 Running CASPER 4 streamlit run app.py A browser window will open with: • Live synthetic telemetry • Synthetic vision + map overlays • Clarity/risk engines • Proposal queue + gating • Audit chain (SHA-256 linked) • CRS-1 effects visualization

📜 Determinism & Replay CASPER 4 produces a run metadata hash (rng seed + configs). The engine supports future replay tooling: • Deterministic per-domain RNG • Frozen telemetry snapshots • Audit chain suitable for verification Replay/diff tooling will be added in CASPER 4.1.

🧭 Intended Use CASPER is built for: • autonomy governance research • HCI experiments • interpretability studies • risk/clarity modeling • safety tooling prototyping It is not a flight system and cannot be used for control of real UAVs.

📄 License MIT

🤝 Contributing PRs welcome. Please open issues for: • new non-kinetic autonomy primitives • deterministic replay tooling • visualization improvements • safety framework extensions