How iOS 15 Calculates Walking Steadiness

Apple’s Walking Steadiness feature in iOS 15 uses the built-in motion sensors inside the iPhone to passively monitor and assess a user’s fall risk. By capturing subtle body movements while the user walks with the phone in a pocket, the system leverages custom algorithms to convert raw physical sensor data into actionable mobility metrics. This article breaks down how the phone’s hardware detects movement, the specific metrics it measures, and how iOS determines whether a user's steadiness is classified as OK, Low, or Very Low.

The Sensors Involved

Walking Steadiness relies primarily on two motion sensors integrated directly into the iPhone:

  • The Accelerometer: Measures non-gravitational acceleration along three axes (X, Y, and Z). It detects changes in velocity, impact forces when the foot strikes the ground, and vertical motion during a stride.
  • The 3-Axis Gyroscope: Measures angular velocity and rotational motion. It captures the natural sway, tilt, and rotation of the pelvis and hips during a walking cycle.

These sensors continuously stream raw motion data to the device's motion coprocessor whenever movement is detected, provided the iPhone is kept in a pocket or worn near the waist.

Core Metrics Measured

Using advanced algorithms validated against clinical balance assessments, iOS 15 extracts four primary mobility metrics from the raw accelerometer and gyroscope data:

  1. Walking Speed: Measures the user’s horizontal speed across flat surfaces. A notable, sustained drop in typical walking speed often correlates with reduced balance.
  2. Step Length: Estimates the physical distance between consecutive steps. Shorter, shuffling steps can indicate compromised stability.
  3. Double Support Time: Measures the percentage of a walking cycle where both feet are touching the ground at the same time. A higher percentage indicates that the user relies more heavily on dual-foot contact to maintain balance.
  4. Walking Asymmetry: Calculates the difference in step timing and dynamics between the left and right foot. Higher asymmetry reflects uneven gait, which increases fall vulnerability.

Algorithmic Evaluation and Risk Classification

The operating system does not evaluate single steps; instead, it aggregates data over extended periods of typical walking across days and weeks to filter out anomalies (such as running, carrying heavy loads, or navigating rough terrain).

Once enough baseline data is gathered, iOS 15 categorizes the user's steadiness into three tiers:

  • OK: Gait metrics fall within normal ranges, indicating low fall risk over the next 12 months.
  • Low: Gait metrics show signs of reduced balance, signaling an increased risk of falling.
  • Very Low: Mobility metrics reflect significant instability, indicating a high risk of falling.

On-Device Processing and Privacy

All sensor processing and machine-learning calculations happen locally on the iPhone using the device's processor. The sensor streams are translated directly into Health app data without sending raw movement logs to Apple servers, ensuring user privacy while maintaining battery-efficient background tracking.