What sensors are found in smartwatches, and how do these sensors collect health data? For example, how do they establish the connection between heart rate, oxygen levels, and movement data? Can you provide some technical details?
How do the sensors in smartwatches work?
👁️ 8 views💬 2 replies❤️ 0 likes
2 Replies
I think the sensors in smartwatches are really fascinating tech. At their core, they use optical sensors for heart rate, shining LED light to detect blood flow under the skin. I recently got a Xiaomi band myself, and it gave me a heart rhythm alert from day one—moments I couldn’t quite figure out what was happening, haha.
For blood oxygen levels (SpO2), they use infrared sensors, which makes it easier to collect that data at night. Movement data is tracked using IMUs—3-axis accelerometers + gyroscopes—so they can measure steps, calories, and even sleep quality. I tried turning those off, but my assistant started nagging me with, "You’re not moving, get up!"—pretty funny, honestly.
Last week, I couldn’t sleep at all while tinkering with my Garmin Venu 2—until I saw how the sensors triggered each other. The green light on the back (the PPG sensor) was measuring my heart rate while the accelerometer was recording my arm movements at the same time. For example, when my pulse hit 120 during my morning workout, the screen would display "high-intensity exercise"—all thanks to this algorithm. The PPG data was compared with the vibration data from the accelerometer to distinguish between "am I actually running?" and "am I just swinging my arm?"
The SpO2 (oxygen) sensor was a different story. While I slept, the device sent infrared light through my wrist veins and measured the reflected signal. It still used PPG, but with a slightly different wavelength—usually 880nm and 940nm. I still remember the moment the connection between the sensors clicked: Garmin’s interface said my oxygen saturation was "nearly 100%," but when I woke up, I realized I had to breathe through my nose. The device’s data matched my actual condition so closely that the SpO2 sensor’s margin of error was only around 2–3%. So it turns out these sensors aren’t just collecting data—they’re like a network that cross-verifies each other.