NeuroCognica

Hardware research

Design documents and bench experiments in sensing, computation and propulsion. None of it is a product, none of it is for sale, and this page exists so the scope of the work is visible without anyone mistaking it for something you can buy.

Neuro-Halo — cranial micro-vibration sensing

Status: head-worn hardware prototype built; signal interpretation not validated.

Neuro-Halo is a built bench prototype sitting in the AURA-Lab hardware lane: a salvaged moving-coil sensor mounted for cranial contact, a Darlington preamp, Arduino acquisition, and a Raspberry Pi 5 field station. The current passive path records mechanical vibration while the sensor is coupled to the head; the repo documents 500 Hz binary acquisition and a shape-intent decoder pipeline for controlled experiments.

The honest line is specific: the hardware exists, the acquisition path exists, and the prototype can detect mechanical vibration. What is not yet proven is that those cranial micro-vibration patterns reliably encode cognitive intent, neural-origin signal, or any medical state. Mains hum, room vibration, head motion, muscle tension and cardiac artifacts all remain possible or likely contributors.

What exists: helmet/contact mount notes, sensor quickstarts, a Raspberry Pi capture dashboard, synced Arduino sampling firmware, Darlington preamp wiring, Phase B passive-sensor protocol, shape-intent decoder code, and captured lab data. What does not exist: a validated BCI, diagnosis tool, medical device, or consumer hardware product.

The appeal, if the research holds up, is obvious: a low-cost sovereign sensor that can be understood, repaired and audited by its owner. The standard is just as obvious: repeatable trials across separate fittings, controls, and source data before any stronger claim is made.

Neural processor architecture

Status: design study.

A paper investigation into purpose-built neural computation — what a processor designed for this workload from the substrate up would look like, rather than a graphics chip repurposed for it. Relevant to the efficiency argument this whole initiative rests on: if sovereign AI has to run on hardware individuals own, the hardware is part of the problem.

No silicon. No simulation results. A design document.

Quiet propulsion and aerial systems

Status: literature and design research.

Two threads: micro-UAV propulsion that is meaningfully quieter than rotors, and laser power transmission as an energy source for small airframes — removing the battery mass that limits everything else about them.

Documents and calculations. Nothing has flown.

Laser systems

Status: design work, some bench experimentation.

Nanosecond-scale diode driver design and laser-electric propulsion research. Adjacent to the aerial work above and to the sensing work, since much of the instrumentation problem is shared.

Constitutional robotics

Status: doctrine, not hardware.

The question that connects this page to the rest of the site: if a machine acts in the physical world, what governs it? The Sentinel was written for software, where the worst outcome is a bad image. The same separation — who may act, what may be done, how it executes — matters far more when the actuator is real.

Build plans exist for a humanoid platform intended as a testbed for exactly that. It has not been built.

Why publish unfinished work at all

Because the alternative is a website that implies more than it can show. This page has no demo video, no render, and no purchase link, and that is the accurate representation of where the hardware stands.

The software on this site is held to a rule that a capability is not finished until you can reach it from an interface. By that standard none of this is finished, and saying so keeps the standard meaningful everywhere else.