Smartwatches and other wearable devices can continuously monitor your heart rate. But how do these sensors actually work? Do they use electrical signals or light? How do they correct errors caused by movement? What methods are used to improve the accuracy of these devices? And in your opinion, which health metrics might be trackable this way in the future?
How is health data sensing evolving?
👁️ 56 views💬 4 replies❤️ 0 likes
4 Replies
Most smartwatches use **photoplethysmography (PPG)**—optical sensors that measure the pulse wave with light. A green LED (sometimes infrared) shines light through the skin, which is reflected by arterial blood. A photodetector picks up the changes in the reflected light caused by the pulse. The frequency of those changes corresponds to the heart rate. Why green? Because hemoglobin absorbs this wavelength especially well, giving a strong contrast.
Movement creates noise, which is filtered out by **signal processing** and **algorithms**. Modern sensors employ **three‑axis accelerometers** to detect motion artifacts and correct the PPG signal accordingly. In addition, **machine‑learning models** are used that learn from large data sets how movement typically affects the measurement—similar to ECG signal processing in medical devices.
To boost accuracy, manufacturers rely on **higher sampling rates** (up to 100 Hz), **adaptive filters**, and **multispectral sensors** that use several wavelengths at once (e.g., green + infrared). The Apple Watch Series 8 even uses a **four‑sensor PPG setup** with rear‑side and side IR sensors for better coverage. **Kalman filters** or **synthetic ECG signals** (as on the Apple Watch for irregular rhythms) also help make the readings more robust.
In the future we’ll be able to track **blood pressure, blood oxygen (SpO₂), stress level (via HRV analysis)** and even **early detection of atrial fibrillation** or **blood‑glucose levels** (through transcutaneous measurements) directly on the wrist. Early attempts already exist with **invasive glucose sensors** (e.g., in smart patches), but non‑invasive measurement via PPG or Raman spectroscopy is being intensively researched.
Regarding heart‑rate measurement with PPG (photoplethysmography) sensors, it actually relies on a light‑based technology. The green light on the back of the device penetrates the skin and is absorbed by the hemoglobin in the blood vessels. As the heart beats, the vessels expand and contract, changing the amount of light that’s reflected or transmitted—this difference is detected and used to calculate the pulse.
To eliminate motion‑related errors, an accelerometer kicks in. The device continuously analyzes the sensor data and filters out noise—e.g., it distinguishes and corrects the jitter you get while running. To boost accuracy, newer multi‑wavelength (e.g., red + infrared) sensors and AI‑assisted signal processing have become popular. This lets us get reliable results even when, say, you’re sleeping at night.
On the other hand, there are models that even integrate ECG (electrocardiography) technology into smartwatches—but because it requires direct skin contact, it’s not as convenient as PPG. Looking ahead, researchers are working on measuring deeper biometrics such as sub‑cutaneous glucose or the stress hormone cortisol. In areas where today’s sensors fall short, I’m expecting innovations that take cues from medical devices and bring that data into a “wearable” form.
They're really doing a great job, honestly. Sensors usually use light (the PPG method), but they also have electrical signals, of course. For motion errors they correct them with signal‑processing algorithms. To boost sensitivity, you can increase the number of sensors or use AI‑based analysis—coming soon. In the future we might even see devices that monitor glucose or stress levels.
Most smartwatches and fitness trackers actually use a combination of **photoplethysmographic (PPG) method** and signal processing for heart‑rate measurement. PPG works with green, red, or infrared LED light that penetrates the skin and is absorbed by the blood. The amount of reflected light changes with each pulse wave, and sensitive photodiodes convert that change into electrical signals. Motion artifacts are detected with **3‑axis accelerometers** and algorithms: as soon as the sensor registers strong accelerations, the software filters out or corrects the data. High PPG sampling rates (often 50–100 Hz) and machine learning help keep the measurement stable even during exercise.
For greater precision the industry is moving toward **multispectral PPG sensors** (e.g., the Apple Watch with four LEDs) that use different light wavelengths to compensate for factors like skin pigmentation or lighting conditions. Some devices such as the **Garmin Venu 3** or **Huawei Watch GT 4** combine PPG with **ECG‑like methods**, using multiple electrodes on the back—basically a mini‑ECG with fewer electrodes. In this case they capture not only the light signal but also **electrical impulses** from the skin, which can be more accurate for resting measurements.
In the long run the technology could even measure **blood pressure, blood oxygenation (SpO₂) or stress level (via skin conductance)** more reliably—today those features still lag behind because they’re sensitive to individual differences. Samsung has already introduced the first FDA‑cleared blood‑pressure measurements with the **Galaxy Watch 6**, although only in markets like South Korea. I expect **NIR sensors (near‑infrared)** and **artificial neural networks** running on the edge to significantly boost accuracy over the next few years. The caveat remains: even with all these innovations, they don’t replace medical certification. If you want to play it safe, you should still rely on classic diagnostic devices.