Building Intelligence Platform

Infrastructure that
understands its
environment.

HomeSpace integrates cyber-physical monitoring, edge compute, and predictive analytics into a unified layer — giving buildings the ability to sense, reason, and respond autonomously.

99.97%
Platform uptime
< 4ms
Edge inference latency
24
Active nodes deployed
847d
Longest continuous uptime
SENSOR MESH · EU-CENTRAL
24 / 24 nodes active
last sync 0s ago
Cyber-Physical Orchestration Edge-Native Sensor Fabric Digital Twin Protocol Predictive Asset Degradation Neuromorphic Inference Zero-Trust Building Mesh Carbon-Aware HVAC Logic Ambient Intelligence Nodes Autonomous Fault Detection Phygital Infrastructure Layer Cyber-Physical Orchestration Edge-Native Sensor Fabric Digital Twin Protocol Predictive Asset Degradation Neuromorphic Inference Zero-Trust Building Mesh Carbon-Aware HVAC Logic Ambient Intelligence Nodes Autonomous Fault Detection Phygital Infrastructure Layer
99.97%
Platform uptime
trailing 12 months
< 4ms
Edge inference
latency
24
Active sensor
nodes deployed
847d
Longest continuous
node uptime

// how it works

From raw signal
to autonomous action.

Every building event travels through five deterministic processing stages — from physical sensor input to real-time actuation — with no manual steps and no cloud dependency for critical paths.

01
📡
Sense
Distributed sensor nodes sample temperature, CO₂, humidity, VOCs, and particulates. Each node runs ESPHome firmware on ESP32 with I²C sensor bus.
BME688 · SCD41 · SPS30
02
⚡
Filter
Raw sensor data is validated, denoised, and timestamped on-device. Outlier rejection runs locally — corrupt readings never reach the network.
< 2ms on-device · Kalman filter
03
📶
Transmit
Filtered payloads are transmitted over LoRaWAN via RAK3172 modules to TTN eu1. MQTT bridge forwards data to the platform core over encrypted WireGuard tunnel.
LoRaWAN SF7 · TTN eu1 · MQTT
04
🧠
Process
Home Assistant core ingests streams into InfluxDB. Node-RED runs automation logic. ML anomaly scoring executes on edge nodes — inference stays local, latency stays sub-4ms.
InfluxDB · Node-RED · HA Core
05
⚙️
Act
Actuation commands are dispatched to HVAC controllers, lighting circuits, and access systems via PWM, Modbus, and Zigbee. Digital twin state updates simultaneously.
PWM · Modbus · Zigbee · Twin sync
Pipeline latency breakdown end-to-end · ms
Sensor sampling0.8 ms
On-device filtering1.9 ms
LoRaWAN uplink28 ms
Platform ingestion3.2 ms
Edge inference3.8 ms
Actuation dispatch2.1 ms
Total critical-path latency ≤ 40 ms
System capability index scored 0–100
Environmental sensing resolution97
Fault detection precision91
HVAC optimisation efficiency94
Digital twin sync fidelity99
Security posture rating96
Carbon reduction vs baseline78
Composite platform score 92.5 / 100

// live telemetry

Real data.
Continuous signal.

Every sensor node streams environmental telemetry at sub-100ms intervals. The platform surfaces anomalies, trends, and efficiency deltas in real time — without manual intervention or polling.

Values update every 10 seconds from live sensor fabric
Sparklines show 60-point rolling window
Anomaly scoring runs on edge — no cloud dependency
Ambient Temperature
BME688 · Zone A
21.4 °C
nominal
CO₂ Concentration
SCD41 · Zone A
412 ppm
good air quality
HVAC Efficiency
PWM · Carbon-aware mode
94.1%
optimal
Relative Humidity
BME688 · Zone A
47.2%
within range
Anomaly Index
Edge inference · local
0.002
clear
LoRaWAN Signal
RAK3172 · TTN eu1
-87 dBm
SF7 / 125kHz

// core modules

Six subsystems.
One coherent layer.

Each module independently deployable but designed to operate as an integrated stack — from raw sensor input to autonomous building response.

🌡️

Ambient Intelligence

Multi-parameter environmental sensing — temperature, humidity, CO₂, VOC, particulates — sampled continuously at sub-100ms intervals.

⚙️

Predictive Maintenance

Anomaly detection on time-series sensor streams. Fault signatures identified early — before failure propagates to critical systems.

🔄

Digital Twin Sync

Every monitored asset maintains a live virtual counterpart. Sub-second state sync enables remote modelling and intervention at any time.

⚡

Edge Compute Layer

Critical inference runs locally on constrained edge hardware. No cloud dependency for time-sensitive decisions. Air-gap capable.

🔒

Zero-Trust Access

Identity-aware, perimeter-free security across all building access vectors. Cryptographically signed audit trail on every event.

🌿

Carbon-Aware Control

Live grid carbon intensity drives autonomous HVAC scheduling and load shifting. ESG compliance generated as a byproduct of operation.