One dashboard, now including what the space is doing.

Count entries/exits using directional detection with hourly traffic and conversion ratio reporting

Track dwell time per zone with rolling 1-hour window heatmap and top zones reporting

Detect brief customer interactions near zones — short dwell (5-30s) indicates product engagement

Track table occupancy cycles — sit-down, meal duration, departure, and turnover rate

Count people by signal disruption intensity with online linear regression calibration

Learn daily usage patterns with hourly Welford statistics, detect wasted energy and report savings estimates

Arrival/departure detection with debounce timers for HVAC control and schedule pattern tracking

Movement-based zone lighting control with auto-off timeout as people move between rooms

Binary occupied/free status with confidence, meeting duration, and utilization tracking

Online k-means codebook of normal signal patterns with anomaly flagging at >2.5 sigma

Baseline learning + combined z-score alerting for unauthorized presence

State machine that detects stationary presence beyond a configurable time threshold

Detect double-presence pattern in doorway monitoring to catch unauthorized entry

Detect concealed metal objects on a person via signal absorption pattern analysis

Poll peer seeds for health status and aggregate cluster-wide metrics

Distribute vectors across seeds with hash partitioning and fan-out query

Manage ESP32 sensor nodes via UDP CSI and serial proxy

Replicate local vector store to peer seed with lag tracking and failure recovery

Multi-seed voting protocol with simple 2-phase commit for coordinated decisions

Bi-directional vector sync between seeds with epoch tracking — only sends changes

Install or remove cogs on peer seeds via apps API fan-out

Distribute queries across seeds based on load metrics so no single one overloads

Discover peer seeds via subnet scan and build mesh topology registry

Share and cross-attest witness chain entries between seeds for tamper detection

Cross-correlation and phase coherence between sensor channels to detect synchronized breathing

HRV analysis using RMSSD and pNN50 from bandpass-filtered heart rate signal for arrhythmia detection

Sleep stage classification (deep, REM, light, wake) from movement and breathing patterns

Walking cadence, symmetry ratio, regularity index, and fall risk scoring via autocorrelation

Combined vital signs dashboard with heart rate, breathing, presence, and apnea detection in one process

Detect seizures via high-frequency burst energy ratio and post-ictal suppression analysis

Detect breathing cessation during sleep via variance spike analysis and CUSUM drift detection

Tracks breathing and heart rate trends using bandpass filtering and zero-crossing BPM

Strange attractor analysis via delay embedding, Lyapunov exponent estimation, and chaos detection

Dynamic Time Warping gesture recognition — record templates and match new signals with confidence scoring

Elastic Weight Consolidation to prevent catastrophic forgetting of learned baselines

Goal-Oriented Action Planning for autonomous action sequence planning and execution

Market data analysis with OHLCV embedding, pattern matching, and trend detection

Brain-inspired leaky integrate-and-fire spiking neural network for movement detection

LTL-like temporal logic monitoring to enforce safety rules on live event streams

Arousal/valence estimation from body movement patterns using the Russell circumplex model

Recognize predefined gesture patterns using Dynamic Time Warping against stored templates

Detect unexplained environmental anomalies via multi-channel correlation breakdown

Composite well-being score from movement, breathing regularity, and activity on a 0-100 scale

Map signal vectors into the Poincare ball model for hierarchical relationship analysis

ESP32 CSI to 17-keypoint skeleton proxy via Hampel outlier rejection and phase sanitization

Detect repeating temporal symmetry patterns via multi-lag autocorrelation analysis

What is running, what it does in plain words, and an audit trail of why it ran

Device health, connectivity, occupancy, alerts and energy at a glance

Provisioning, signal quality, firmware, diagnostics and consented remote support

What is collected, where it is processed, who can see it and for how long

Presence, falls, unusual motion, environmental sensors and escalation state

Detect tampered/spoofed signals via statistical tests for artificial regularity and temporal inconsistencies

Signal quality monitor with per-channel coherence metrics, quality scores, and noise floor estimation

Core gesture recognition building block using template-based DTW matching

Grover-inspired amplitude amplification for rapid anomaly detection in large feature spaces

Sensor mesh health monitor with dropped node detection and automatic redistribution

Enforce maximum headcount in controlled environments via simultaneous presence counting

Monitor workers in confined spaces with presence duration tracking, breathing analysis, and inactivity alerts

Detect large moving objects via strong multi-channel signal disruption with rapid approach alerting

Detect animal activity patterns, distress signals, and escape events via CSI analysis

Detect dangerous vibrations in buildings or machines via high-pass filter and RMS energy analysis

Filter signals by channel-pair coherence, only passing clean data above threshold with SNR estimation

Adaptive sensing that focuses processing on highest-information channels by variance and entropy ranking

Estimate missing channel data via regression from correlated channels

Downsample old data to save memory while keeping recent data at full resolution