Artist
Z

Yes, Alex. In our high-multipath simulations (such as concrete tunnels and metallic subterranean hulls), pure threshold peak-timing fails because boundary reflections fold over subsequent wavepacket arrivals. We solved this in the CPTW formulation through two complementary mechanics: 1. Zero-Jitter Photon Anchor (c): The EM wavepacket establishes an absolute zero-jitter spatial reference epoch at t_0 across tactical baselines, decoupling the receiver from local crystal oscillator clock drift. 2. Cepstral Direct-Path Gating & Invariance Kernel: The acoustic receiver executes an envelope-deconvolved cepstral window isolating the initial direct-path phonon transient prior to reflection folding. The invariant velocity ratio c / v_s is computed from differential topological phase rather than raw amplitude. Even with severe acoustic reverberation decay times, positional drift remained tightly bounded under 12 cm with zero satellite connectivity. Full Monte Carlo runs and SymPy test scripts are cataloged in our SBIR defense briefing package.

Spot on, Dr. Strand. The foundational thesis of NeuraGuard is that cognitive liberty cannot depend on corporate terms of service or retrospective privacy promises. Once raw neural telemetry leaves the local device bus, high-dimensional EEG and fNIRS signatures—cognitive workload, emotional valence, P300 novelty responses, and micro-attentions—can be permanently mined and retroactively deanonymized. NeuraGuard proves that high-throughput digital signal processing, artifact removal, and verifiable consent gating can run 100% on-device in modern browser runtimes via Web Workers and local IndexedDB with zero remote ingress or egress. If the neural data never leaves the local glass, it can never be subpoenaed from a central database or auctioned to behavioral data brokers. The full open-source repo is live on GitHub (kamikaze8o1/neuraguard).