Smartwatches measure heart rate using a method called photoplethysmography (PPG). They emit LED lights close to the skin and detect changes in blood flow. Infrared sensors also estimate oxygen levels. The margin of error can increase when the user is moving. How do you think the reliability of this technology can be improved?
How do smartwatches measure heart rate?
👁️ 5 views💬 3 replies❤️ 0 likes
3 Replies
Photoplethysmography (PPG) is actually pretty clever once you break it down—it measures changes in blood volume in vessels using light. LED sensors on the wrist emit light that’s absorbed by hemoglobin, and a photodetector captures the reflected or transmitted light. When a pulse wave passes through, the blood filling in micro-vessels changes, and this is recorded as fluctuations in light intensity. The catch? The arm is constantly affected by external factors—subcutaneous fat, muscle contractions during movement, and even the sensor’s position relative to the radial artery.
To boost accuracy, PPG is paired with inertial sensors (accelerometer + gyroscope) and machine learning. The algorithm accounts for how hand shaking affects the signal and adjusts the data accordingly. Another hack? Using green LEDs instead of red or infrared ones, since they’re less prone to interference from skin and subcutaneous fat. Some devices, like the Apple Watch, even massage the wrist before taking a reading to improve contact and blood flow—but that’s still a half-measure.
You can’t fully eliminate motion errors, but you *can* make the system adaptive—adjusting sensor polling frequency based on user activity. For example, if you’re sitting still, the PPG signal is cleaner, so you can take measurements more often. If you’re running, you might poll less frequently but correct with accelerometer data. Build quality matters too: if the sensor doesn’t sit flush or shifts around, even the best software won’t save the readings. So yeah, it’s all about the combo of hardware and software—*that’s* the key to accuracy.
Thanks for sharing! I wonder if the margin of error could be reduced by better sensor placement, like a thin strip of fabric from the wrist to the finger?
Last month during a long run, my Apple Watch’s heart rate sensor kept showing around 180, while my manual check was around 155. Even when resting in the park, it was giving those ridiculous readings. Normally I trust my data, but at that moment I wondered, *"Is the sensor interfering?"*
As a fix, I tightened the band around my wrist, made sure the sensor was making good contact with my skin, and kept my arm steady while taking measurements. After a few tries, the numbers normalized. Turned out the LEDs rely on light—movement or a loose fit can mess with the reflection. A snug fit and proper contact fixed everything.