Wearables and Smart Tech

How Fitness Trackers Measure Steps and Heart Rate

See how accelerometers estimate steps and how light-based sensors estimate heart rate in fitness bands and watches, and why the numbers are never exact.

On this page
  1. Estimates, not direct measurements
  2. How steps are counted
  3. How heart rate is estimated
  4. Other sensors you may find
  5. Getting more out of the numbers
  6. Health caution
  7. What happens to the data

Estimates, not direct measurements

A fitness tracker does not literally see your steps or count your heartbeats the way a person would. It collects signals from small sensors and uses software to guess what they mean. Understanding that difference explains most of the surprises people have with their devices, such as a step total that rises while they are driving or a heart-rate reading that looks too high during a workout.

How steps are counted

The accelerometer

The key part is an accelerometer, a tiny sensor that measures changes in speed along three directions. Each time you walk, your body rises and falls and your arm swings, and this creates a repeating pattern in the data. The tracker’s software looks for peaks in that pattern that match the rhythm and size of a typical step.

Many devices also include a gyroscope, which senses rotation, and the two together help distinguish a swinging walk from other movements. These components are described more broadly in the article on what device sensors do.

Why step counts drift

Because the software is matching patterns, it can be fooled:

  • Arm movements without walking, such as brushing teeth or stirring a pot, may produce false steps.
  • Pushing a stroller or shopping cart keeps the wrist still, so steps may be missed.
  • Slow walking produces weaker signals that may be filtered out.
  • Bumps in a car or on a train can occasionally register as steps.

Where the tracker is worn matters too. Some are worn at the hip or on a shoe, which can produce a different pattern than the wrist.

Most trackers estimate distance by multiplying steps by a stride length based on the height and sex you enter, or by learning your stride from outdoor walks where satellite positioning is available. The result is an approximation.

How heart rate is estimated

Optical sensing

On the back of most wrist devices are small lights, usually green, with a light sensor next to them. The method is called photoplethysmography. Blood absorbs some of the light, and each pulse slightly changes how much is absorbed. The sensor reads the rapid rise and fall in reflected light, and the software counts the pulses per minute.

Why optical readings can be unreliable

The signal is weak and easily disturbed:

  • Motion. Vigorous arm movement shakes the sensor against the skin and adds noise. Activities with gripping or wrist bending, such as weightlifting, are often harder for these sensors than steady walking or cycling.
  • Fit. A loose band lets in stray light and shifts position. A snug but comfortable fit usually helps.
  • Skin and environment. Cold skin with reduced blood flow, dense hair, darker skin tones and tattoos can all affect how light behaves, though makers say they try to account for this.
  • Delay. Readings are often smoothed over several seconds, so quick changes may appear late.

Heart-rate chest straps use a different method that picks up the heart’s electrical signal. They are generally considered more reliable during intense exercise, though they too can give odd readings if the contacts are dry.

Derived numbers

Many other figures are calculated rather than measured. Resting heart rate, estimated calories burned, workout intensity zones and sleep stages come from models that combine heart rate, movement and personal details. Calorie estimates in particular can be off by a wide margin, since they depend on assumptions about your body. Some use techniques related to machine learning basics, in which software is trained on examples to classify activity or sleep, but a trained model is still making an educated guess.

Other sensors you may find

Depending on the model, a tracker may also include:

  • A barometer to count floors climbed by sensing air pressure changes.
  • A skin temperature sensor, mainly useful for noticing trends.
  • A blood oxygen estimate, using red and infrared light. These values are typically presented as wellness information, not as a clinical reading.
  • Satellite positioning for outdoor routes, which uses more battery, a trade-off familiar from battery drain on phones.

Getting more out of the numbers

  • Look at trends over days and weeks rather than any single reading.
  • Compare like with like. A tracker is usually consistent with itself even if it differs from another brand.
  • Wear it snugly, a finger or so above the wrist bone, for workouts.
  • Keep the app and device updated, since manufacturers adjust their algorithms through software updates.
  • Enter your height, weight and age accurately, since estimates depend on them.

Health caution

A tracker can encourage activity and flag changes worth discussing, but it is not a diagnostic tool unless it is specifically a regulated medical device. An unusual reading, such as a very fast or irregular pulse together with symptoms like dizziness or chest discomfort, calls for professional medical advice rather than a second look at the app. Likewise, a normal reading does not guarantee good health.

What happens to the data

Trackers store a detailed record of movement, sleep and heart rate, and this information is typically synced to the maker’s cloud service. It can reveal daily routines and health patterns, which makes it sensitive. Checking the app’s privacy policies, limiting what is shared with other apps, and understanding whether data can be deleted are sensible steps. Wider questions about who may handle such information are discussed under digital privacy.

The complete guide to wearables and smart tech

Wearable Technology Explained
Read the guide

Keep exploring

Guides from other topics

All topics →