How does Microtrace Work?
Traditional drinking fountains operate blindly after installation. MicroTrace transforms the fountain into a self-verifying system by layering sensing → physical state estimation → AI intelligence → spatial insight around a UV-C disinfection core.
End to End Flow
A step-by-step overview of the process from when water enters the system to when it exits
#1
Entry
Water enters MicroTrace through an offset tube that causes the water to spiral and spend more time in the disinfecting chamber
#2
Sensing
Sensors continuously measure UV intensity, flow, pressure, turbidity, and temperature
#3
Kalman Filter
Kalman filter fuses data into accurate physical state estimates
#4
UV Analysis
UV dose is verified against microbial inactivation thresholds
#5
AI Layer
AI evaluates short-term safety and long-term trends
#6
Feedback Loop
System status updates in real time + UV Adjustment
#7
Data Collection
Data contributes to facility-wide and geospatial insights
Sensor Layer: Measuring System Behavior (Raw Observations)
MicroTrace integrates a compact, low-power sensor suite that continuously monitors the physical conditions that determine whether UV disinfection is actually effective.
Kalman Filter Layer: Accurate Physical State Estimation
A proven state-estimation algorithm used in aerospace, robotics, and autonomous systems
What the Kalman Filter Does
The Kalman filter:
-
Fuses multiple noisy sensor inputs
-
Incorporates physical flow and UV exposure models
-
Estimates hidden system states that cannot be directly measured
Estimated States (What Actually Matters)
Instead of trusting raw sensors, MicroTrace estimates:
-
True flow velocity
-
Actual residence time
-
Delivered UV dose
-
System health indicators (lamp aging, clogging trends)
Why This Matters
Microbial inactivation depends on UV dose = intensity × exposure time.
The Kalman filter ensures this calculation reflects real system behavior, not momentary sensor noise.
It also:
-
Corrects for sensor drift
-
Smooths transient fluctuations
-
Detects slow degradation trends long before failure
This layer converts data into physics-grounded truth.
AI Layer: Predictive Intelligence
In addition to stabilized system estimates, MicroTrace applies a transformer-based AI model trained on time-series sensor data.
What the AI Model Analyzes
-
UV dose trends
-
Flow variability patterns
-
Pressure and turbidity changes over time
-
Temperature-adjusted effectiveness
What the AI Predicts
-
Contamination risk level: low / moderate / high
-
Early UV lamp degradation
-
Flow instability or abnormal usage
-
Filter clogging or biofilm growth patterns
Why AI Is Necessary
Not all failures are abrupt. Many are slow, nonlinear, and context-dependent.
The AI distinguishes:
-
Normal daily fluctuations
-
vs. meaningful deviations that indicate risk
This allows preventive intervention, not reactive response.
Geospatial Heatmap: System-Level Insight
Each MicroTrace unit logs anonymized, location-tagged safety data.
How It Works
-
Latitude and longitude are included as model inputs
-
The AI learns spatial correlations across buildings, campuses, or cities
-
Contamination likelihood is predicted across locations
Output
-
Geospatial heatmaps showing risk intensity
-
Identification of systemic plumbing issues
-
Prioritization of maintenance and upgrades
This turns isolated fountains into a public-health monitoring network.