Google researchers have demonstrated that consumer smartwatches can accurately estimate advanced walking metrics like speed and step length. A large-scale study with 246 participants showed the wrist-worn devices performed comparably to smartphones, despite smartphones having more training data. This bridges a gap where smartwatches were historically less reliable for comprehensive gait analysis.

The study used a deep learning model on raw accelerometer and gyroscope data from a Pixel Watch, incorporating user height to improve accuracy. Researchers compared the smartwatch against lab-grade gait walkways and smartphone data from various pocket placements. The findings suggest smartwatches offer a practical, consistent platform for continuous tracking, free from smartphone placement constraints.

This work moves gait analysis from expensive labs into everyday life. Continuous monitoring could enable early detection of neurological conditions, fall risks, and personalized rehabilitation. Future applications may allow for longitudinal health tracking outside clinical settings, making advanced biomechanical insights more accessible to consumers and clinicians alike.