A clinical-grade 905nm LiDAR voxel tracker with hardware-enforced human sovereignty. The kinetic arm scans tissue without contact. The Sacred Pause gate ensures no output propagates without human authorisation. The 1.1x surplus guarantees the system never degrades.
Click the glowing dots on the device to explore each subsystem.
905nm
LiDAR wavelength
<100ms
Sacred Pause
1.1x
Compute surplus
57.6M
Max voxels
SPEC 1
NEC-TS-NIVT-905
905nm LiDAR Voxel Tracker
The kinetic arm houses a pulsed 905nm InGaAs laser, MEMS micro-mirror scanner, and 256×256 SPAD detector array. It performs time-of-flight photon detection to build volumetric tissue maps at 200μm lateral resolution without physical contact, ionizing radiation, or contrast agents.
Wavelength
905 ± 5nm
InGaAs pulsed diode
Detector
256×256
SPAD array · 10μm pitch
Voxel Size
200×200×50μm
Up to 57.6M voxels
Scan Rate
10 Hz
MEMS 2-axis raster
Penetration
2–8mm
Tissue-dependent
Safety
Class 1
IEC 60825-1 eye-safe
Physical location: Sensor head at top of kinetic arm. MEMS mirror in arm shaft. Signal processing routed to FPGA/GPU in main body via shielded flex cable.
SPEC 2
NEC-TS-ACI-SP1
Sacred Pause Gate
A hardware-enforced timing gate burned into the FPGA fabric on a dedicated clock region with an independent 100kHz oscillator. Every kinetic output passes through this gate. The gate does not evaluate. It waits. No software override exists. A broken gate welds shut.
Hold-Off Range
1–99.9ms
DIP switch configured
Gate Resolution
10μs
100kHz independent clock
Bypass Path
NONE
OTP fuse · AES-256 bitstream
Gate Points
7 gates
ACI-G01 through G07
Auth Input
Dual-button
+ pedal, touch, biometric
Failure Mode
Inaction
Every failure → safe state
Physical location: Dedicated FPGA clock region in main body, isolated from processing pipeline. Auth input port on right panel. E-stop on top face (hardwired, bypasses gate via direct power cutoff).
SPEC 3
NEC-TS-SIS-110
1.1x Sovereign Infrastructure Surplus
Every resource domain — compute, memory, storage, network — is provisioned at 1.1× worst-case peak demand. If any domain falls below its Sovereign Floor, the system shuts down rather than degrade. There is no degraded mode. The SHI LED strip on the device body shows real-time headroom status.
FPGA Headroom
≥ 13%
Logic utilisation ceiling 87%
GPU Headroom
≥ 18%
CUDA peak 82%
Memory BW
≥ 42.2 GB/s
LPDDR5 dual-channel + ECC
Thermal Policy
NO THROTTLE
N+1 fans · ≤85°C junction
Sovereign Floor
D_peak/1.1
Breach → immediate shutdown
SHI Indicator
0.00–1.00
10ms refresh · LED + display
Physical location: Distributed across main body. FPGA + GPU centre-mounted. LPDDR5 flanking compute. NVMe below. N+1 cooling fans at base. SHI status LED strip on left panel. WORM audit chip on right panel.
Signal Flow · Kinetic Arm to Human Hand
📡
SCAN
905nm LiDAR
Photons → Voxels
→
⚡
PROCESS
1.1x Surplus
FPGA + GPU pipeline
→
⏸
PAUSE
Sacred Gate
System halts. Waits.
→
✋
HUMAN
The Hand
Intent → Execution
Physical Specifications
180mm
Height (body)
120mm
Width (body)
80mm
Depth (body)
≤1.2kg
Sensor Head
≤45W
Total Power
🔦
AI ROLE
AI as Flashlight — Illuminating, Never Deciding
AI functions strictly as a flashlight — illuminating risks and patterns for the Human Pilot. It does not steer. It does not choose. It does not recommend. The clinician is the decision maker. AI is a Humble AI — a sterile instrument that serves human sovereignty.
HUMBLE AI
FLASHLIGHT ONLY
HUMAN PILOT
NON-AGENTIC
🫥 LiDAR Point Cloud · Fall Detection Geometry · Privacy First
When a person-shaped volume transitions from vertical to horizontal, the system detects the geometry change — not the person. No cameras. No faces. No identity. Only point-cloud volumes in 3D space. Privacy is not a feature. Privacy is the architecture.
LIDAR POINT CLOUD · HUMAN PRESENCE · 3ZEROS · NO CAMERA
3 PERSONS · ALL UPRIGHT
📊 Live System Dashboard
🔐 Governance Protocols
🪪
Lanyard Protocol
Only Senior Clinicians wearing the Green Lanyard may authorise clinical actions. Junior staff are counter-intuitively excluded — protecting both patients and trainees from AI-assisted overreach.
SENIOR ACCESS ONLY
⏸
Sacred Pause Gate
FPGA-enforced timing gate. Every output freezes. The system does not evaluate — it waits. No software override exists. A broken gate welds shut. Human verification mandatory before any action proceeds.
FPGA ARMED
🔴
Red Code Protocol
AI failure triggers mandatory 24-hour pause for root cause analysis. Manual override physically accessible. System cannot self-restart. Human hands must physically re-engage the device.
STANDBY
ALERT STATUS
RED CODE PROTOCOL
DIAGNOSTIC
23:59:49
On AI failure: system halts, 24-hour mandatory pause, root cause analysis required. Manual Override physically accessible. The human pilot remains in command at all times.
AI HALT
SIG FREEZE
ROOT CAUSE
MANUAL OVR
HUMAN RE-ENGAGE
Clinical Filter · Zero-Trust · Non-Agentic
Nightingale's Eyes Care™ as Clinical Filter
ZERO-TRUST NON-AGENTIC SAFETY LAYER — VALIDATES EVERY AI OUTPUT BEFORE IT TOUCHES A PATIENT, RECORD, OR WORKFLOW
A non-agentic circuit breaker that sits between every agentic AI system and the patient. It filters. It scores. It escalates. It never acts. It never prescribes. It never orders.
AI Agent Outputs
All treated as untrusted suggestions
↓ ↓ ↓
👁️
NIGHTINGALE FILTER
ABC+2S+H Guardrails · Medical Grounding · Privacy-by-Physics
✓
Medical Validity — checked against clinical guidelines, formularies, local SOPs
✓
Bias Detection — demographic, diagnostic, and treatment bias screening
✓
PHI Leakage — protected health information containment enforcement
✓
Scope Creep — assistive vs autonomous boundary enforcement
↓ ↓ ↓
Safe
Passes to clinician view
Human Check
Held for clinician review
Block
Rejected & logged
🔒 Zero-Trust AI — Enforce Do No Harm
Untrusted by Default
Every LLM and agent answer is treated as an untrusted suggestion. No confidence threshold grants autonomy. No track record earns trust. Every output is verified.
Fail-Safe to Human
High-risk or anomalous behaviour forces a fail-safe to human review. Automation does not continue. The system waits. The Sacred Pause Gate enforces this in hardware.
Assistive ≠ Autonomous
Distinguishes assistive from autonomous tools. Stricter rules when output could influence diagnosis, orders, triage, or consent. The boundary is absolute.
Continuous Monitoring
Watches for drift, bias, and new failure modes. Supports lifecycle audits as expected in modern medical-AI governance. The filter learns what goes wrong — it never learns to act.
14:29:30AI diagnostic suggestion — bias flag raised on demographic weightingBLOCKED
🏛️ Governance Stack — Where Nightingale Sits
National Policy
IMDA MGF for Agentic AI · sets outer limits for what agents may do
Sovereign
ABC+2S+H Guardian Framework · clinical chain of command, audit, platinum standard
Clinical Filter
Nightingale's Eyes Care™ · filters, scores, escalates — never acts
Local Agents
Hospital IT + vendors · must route through Nightingale + ABC+2S+H rails
NAAI Ecosystem Rule
Every agentic system in the hospital — beds, cameras, EHR copilots, scheduling agents, robotic carts — must route clinical-adjacent actions through Nightingale's filter before they touch patients, staff, or records. No exception. No bypass. The filter is the gate. The clinician holds the key.
WHITE PAPER 1.0 · RISK SCREENS
Three Clinical Risk Screens
Nightingale's Eyes Care™ non-invasively monitors for three high-priority eldercare risks. All outputs pass through the Clinical Filter before any clinician view. The system observes, scores, and escalates — it never diagnoses or acts.
🚶
🚨
Falls
RISK SCREEN 01
905nm LiDAR voxel tracking detects gait deviation, postural instability, and sudden altitude change. Fall events trigger immediate escalation through the Sacred Pause Gate — no autonomous response.
Behavioural geometry: agitation index, nocturnal micro-movement clusters, and postural restlessness patterns. Nightingale's Scan builds a temporal motion fingerprint, flagging drift from baseline without any camera.
24h
BASELINE WINDOW
3ZEROS
PRIVACY COMPLIANT
FILTER PATHWAY
Motion voxel stream → Agitation index engine → Delirium risk tier → Nightingale Filter → Consultant psychiatry flag only
Conceptual draft for internal canon-building within the WISL series; not a deployed mission, not public guidance, and not effective until NLB approval of the first 15 WISL entries. All scan data shown is illustrative for architectural reference only. No patient data is referenced.
🛡️
Sovereign Governance Architecture
PLATINUM STANDARD
The constitutional framework ensuring that no AI system, regardless of capability, may exceed the authority granted by its human sovereign. Governance is not an add-on — it is the foundation upon which every circuit is built.
🐯
SECURED AGAINST HARM
Sacred Legacy of Human Dignity
This technology exists to protect the dignity of those who gave us life. Every circuit, every gate, every pause — built in service of human sovereignty. AI must never replace the warmth of a human hand.