The suit is one closed loop: Observe → Estimate → Encode → Optimize → Verify → Act → Measure. Hover any node for a summary; tap it for full technical spec, physics/math and features. Colour marks the evidence class.
The architecture does not claim to defeat conservation of momentum. Distributed area, ducting, Coandă attachment and vector control are engineering mechanisms for changing how momentum is coupled into the vehicle — not a way around the actuator-disk penalty.
The pilot biomechanical safe set X_bio ⊂ ℝⁿ is a hard constraint of the same rank as any flight-envelope constraint.
Architecture-level indicators, not measured performance.
Attachment length, deflection θ and useful force depend on slot height h_j, radius R_c, jet velocity and local Re/Ma. Separation collapses the negative-Cp lobe and dumps propulsive load onto the pilot.
Thrust-to-power is far below mechanical fans, so EHD electrodes and dielectric-barrier plasma are used to hold Coandă attachment under gust and maneuver — a flow-control technology, never primary lift.
Photon thrust is ruled out as primary lift. Any Casimir / DCE / sonoluminescent thrust claim must satisfy the Ford–Roman quantum inequality before touching flight hardware.
Every subsystem exchanges conserved quantities explicitly: mass → momentum → energy → heat → information. The optional ψ (QED) state is used only by HALO-CST.
Structure, flow, energy, thermal, sensing, compute and human-machine control are optimized jointly, not independently — the core architectural principle inherited from NWO COANDA VTOL.
| Layer | Function |
|---|---|
| Outer skin | aerodynamic shape · abrasion |
| Coandă surface | attach & turn boundary flow |
| Micro/nano channels | distribute plenum air to outlets |
| EHD electrodes | hold attachment under gust |
| Piezoelectric | pressure/vibration sensing + harvest |
| Dielectric | field insulation from body |
| Thermal (graphene/ceramic) | heat spreading + bus routing |
| Structural | load bearing to exoskeleton |
| Distributed fan layer | ≈ 95%+ of useful lifting energy |
Reference: H₂ fuel cell + high-power Li-ion peaking pack. Solar (~400 W), thermoelectric fabric, regenerative joints and micro-harvesting are supplementary layers — every source ultimately draws on a real physical flux; nothing bypasses energy conservation.
Fuel-cell, motor, electronics and body heat feed one bus used for heating, Peltier cooling or thermoelectric regeneration. dS_prod/dt ≥ 0 throughout.
A JARVIS-class reasoning layer plans and explains in natural language, but a certified deterministic safety filter projects every AI intent into the safe operating set before it reaches an actuator.
Voice/gesture/gaze/EMG/EEG interaction; HUD overlays; explanations on demand; robot-fleet command decomposed to ROS2 tasks; biometric fallback triggers autonomous safe-descent on pilot incapacitation.
LIVE · huggingface.co/spaces/CPater/chainstateControl allocation across N fans, MPC trajectory optimization, digital-twin evolution, correlated-equilibrium fleet coordination and sensor-fusion attention all map onto the same energy-landscape IR and dispatch through the CS.PHYSOPT ISA. HALO does not require CHAINSTATE-X to fly — the base configuration uses CS.* CMOS — but a CHAINSTATE-X base station gives one to two orders of magnitude better performance-per-watt on optimization workloads.
Each variant inherits the same exoskeleton, adaptive-skin, energy-bus, CHAINSTATE/safety-kernel and digital-twin core while changing the propulsion and environmental envelope. Every variant is intended flight-worthy for an adult pilot at appropriate technical scale.
| Variant | Ground | Air | High-alt | Water | Vacuum |
|---|---|---|---|---|---|
| HALO-0 | Primary | — | — | — | — |
| HALO-1 | Yes | Tether | — | — | — |
| HALO-2 | Yes | Primary | — | — | — |
| HALO-3 | Yes | Primary | Yes | — | — |
| HALO-4 | Yes | Yes | — | Primary | — |
| HALO-X | Ltd | Yes | Primary | — | — |
| HALO-S | Ltd | — | Ascent | — | Primary |
| HALO-CST | Research configuration across every medium | ||||
A given exit velocity produces ~800× more thrust in water than in air at equal disk area, so cavitation-aware HALO-4 propulsion is engineering-viable at moderate exit velocities.
A 12-fan reference remains hover-capable with up to three simultaneous fan failures; the safety kernel handles thrust reallocation. Critical sensors are triplicated (S_valid = Majority).
A prior draft implied 10–100 MW from sonoluminescence. At the same per-event energy and rate the real figure is ~1.6 mW — off by roughly ten orders of magnitude. HALO-X depends on no such excess.
Recovering 10 MW at that per-event energy would need R ≈ 6.24×10²¹ events/s — not physically plausible for a few-bubble cell. Sonoluminescence is retained as a measurable optical–acoustic diagnostic and a strongly nonlinear reservoir, not an energy source.
Three redundant force channels (torsion balance, interferometric, momentum-recoil) plus dual optical channels (SPAD + spectrometer). No CST measurement is ever routed into flight control.
Only a statistically significant residual survives as evidence for a new interaction. DCE (Wilson 2011) is real, but the photon energy comes from the modulation drive — not from the vacuum.
Industrial exoskeleton → tethered propulsion → free flight → forward flight → transmedium → high-altitude → space, with speculative propulsion kept on a separate falsification branch.
HALO-0 enters this market directly as the near-term revenue engine; each subsequent variant extends into adjacent verticals with higher long-term defensibility. Source: ABI Research 3Q 2025 — verify before investment use.
| Player | HALO asymmetry |
|---|---|
| Gravity Ind. | ships exo + fleet-data under the flight variant |
| German Bionic | same exo logic + certified path to flight |
| Figure 02 | wraps the AI around a human, not replacing one |
| Boston Dyn. | a HALO pilot commands Atlas-class robots |
| Axiom / NASA | HALO-S adopts AxEMU life-support envelope |
| eVTOL fleets | fills the wearable segment they don't address |
An airspace-level traffic manager acts as the correlation device. Every added pilot or robot raises the coordination value of the whole platform — a network effect single-airframe fleets cannot generate.
HALO-0 exoskeleton → data + revenue + manufacturing → HALO-1 tethered → HALO-2 free flight → HALO-3 cruise → HALO-4 transmedium → HALO-X high-altitude → HALO-S space. Every HALO-2+ pilot trains inside a HALO-0 context on the same software stack before going airborne.
RaaS, FaaS and SaaS dominate on defensibility and recurring revenue.
HALO is the wearable-mobility node of the wider NWO ecosystem — deployed as a HuggingFace static Space with settlement anchored on Base mainnet — and every quantitative claim ships with a stated null-result definition.
| Contract | Address |
|---|---|
| NWOIdentityRegistry | 0x78455AFd…De1dAfF8 |
| NWOAccessController | 0x29d177be…459a0f50 |
| NWOPaymentProcessor | 0x4afa4618…3d5abd7c |
| MetaStateSplitter | 0x93a7962f…833b1BE4 · 35/35/30 + 15% affiliate |
| STATE token | 0x9533DF99…5F727d8a |
| USDC (Base) | 0x833589fC…bdA02913 |
Every sensor and action is authenticated against the identity registry and signed by the pilot's Cardiac-SDK ECG soul-bound NFT, giving fleet operators, insurers and regulators a per-suit-hour audit trail. Human pilots settle mission and flight-hour revenue exactly like any NWO robot agent.
| Layer / hypothesis | Falsification predicate | If null |
|---|---|---|
| L1 · hover thrust | measured Tᵢ > 20% below actuator-disk theory | increase area or accept lower T/W |
| L1 · Coandă attachment | C_attach < 0.5 in design envelope | add EHD / plasma control |
| L1 · endurance | t_mission < 60% of Tier-A target | re-tier energy; add peak buffer |
| L2 · HALO-4 drag | no useful drag-reduction band | conventional streamlined swimming |
| L2 · photonic Ising | coupling fidelity < 0.90 at 256 spins | fall back to CS.* CMOS |
| L3 · Casimir excess | F_residual < 5σ, systematics bounded | conventional metrology only |
| L3 · metric engineering | no admissible T_μν for any h_μν | layer abandoned; variants unaffected |
NWO HALO, Volume I (2026) by Ciprian Florin Pater — a peer-review-ready engineering, research and commercialization framework organized under a three-layer evidence discipline.
Layer 1 — conventional core. Distributed Coandă propulsion, exoskeleton, adaptive skin, energy tiers, thermal management and the deterministic safety controller. Newton–Euler / actuator-disk / Navier–Stokes / thermodynamics — every equation used to size HALO-0 → HALO-S is engineering-grade and design-usable today.
Layer 2 — HALO-X research. Cavitation-aware transmedium drag control, sonoluminescence diagnostics, Casimir and dynamical-Casimir metrology, and the CHAINSTATE-X photonic Ising / MZI / reservoir / phononic optimization substrate. Experimentally testable, not assumed energy sources.
Layer 3 — speculative. Metric-engineering and information-geometry extensions, retained only under the observation → reconstruction → residual → hypothesis → falsifiable-theory pipeline. The paper's key correction rebuilds the sonoluminescence energy budget to ~1.6 mW and states the excess-energy falsification test explicitly.
Affiliated: University of Agder, Kristiansand, Norway · NWO Capital / NWO Robotics.
Status · Design / Roadmap