Live demo · wearable signal intelligence

From noisy sensors to a clear read on how someone is doing

Wearable data is messy. Movement, sweat and loose electrodes bury the signal. This demo shows a full pipeline running live in your browser: it cleans an ECG, detects every heartbeat, fuses heart rate variability, breathing, skin conductance and motion, and turns them into a stress and mood estimate that can act on its own.

  • ECG filtering
  • Motion-artifact detection
  • Beat detection
  • HRV
  • Sensor fusion
  • Real-time alerts
Interactive demo · try the controls on the right
Live telemetry · simulated runner

Mira, morning run

Good mood
00:00Easy run · 9.5 km/h
Heart rate--bpm--
HRV · RMSSD--ms--
Breathing--/minchest strap
Skin conductance--µS--
Cadence--spmwrist IMU
Stress index--/100--

ECG rhythm strip

chest patch, lead II · 250 Hz · 25 mm/s · 10 mm/mV Signal quality -- Motion flag --

Amber shading: moments the motion-based detector flags as likely ECG artifact, using the accelerometer as in Halitech's published research. Try Clap / wave arms or Sprint.

Skin conductance

--
wrist EDA · 30 s · ticks = detected bursts

Respiration

--
chest-strap stretch · 30 s

Footstrike

--
vertical accel, g · 8 s

Timeline

alerts, nudges and what actually happened on the route

    Mood estimate

    valence × arousal · last 60 s
    ----

    Run controls

    Pace
    Sensor noise
    45%
    Make something happen
    Monitor

    Have wearable or sensor data that needs this kind of clarity?

    Halitech Solutions builds signal processing and applied AI for wearable and health monitoring, from research prototype to working system.

    Talk to Halitech

    The motion-artifact flags are a simplified illustration of the approach in Dr. Hamidi's peer-reviewed research, which detects ECG artifacts from the device's accelerometer. The published method uses wavelet scattering features and a trained classifier; this demo uses a simple motion-energy rule and is not that model. Mira and every number here are simulated in your browser: a synthetic ECG with realistic noise, a Pan-Tompkins style beat detector, and wearable streams that respond to pace, fatigue and stress. Wearables measure arousal and stress, not mood itself, so the mood estimate is a heuristic model and a real deployment would need the wearer's consent and per-person calibration. Timeline entries marked "Route" are the hidden ground truth, shown so you can see how quickly the monitor reacts. No personal data is collected.

    © 2026 Halitech Solutions Ltd. · Halifax, Nova ScotiaDemo only · not a medical device