Accurate walking and running speed estimation using wrist inertial data.

Area of Science:

• Sports Science
• Biomechanical Engineering
• Wearable Technology

Background:

• Accurate running speed estimation is crucial for training and performance analysis.
• Existing methods like GPS or foot pods have limitations in accuracy or convenience.
• Wearable accelerometry offers a potential solution for unobtrusive speed monitoring.

Purpose of the Study:

• To develop and validate an accelerometry-based device for robust running speed estimation.
• To integrate this technology into a user-friendly, watch-like form factor.
• To compare the device's accuracy against established speed measurement techniques.

Main Methods:

• Utilizing a leg-and-arm dynamic motion model applied to 3D accelerometer signals.
• Implementing a calibration procedure using a known distance or constant speed.
• Testing the device across a range of walking and running speeds (1.8–19.8 km/h) with eleven subjects.

Main Results:

• Achieved unbiased running speed estimations.
• Demonstrated high accuracy with 2nd and 3rd quartiles of relative error within ±5%.
• Results showed performance comparable to traditional foot pod devices.

Conclusions:

• The accelerometry-based watch device provides a robust and accurate method for running speed estimation.
• This technology offers a convenient and effective alternative to existing speed tracking solutions.
• Further validation in diverse running conditions and populations is warranted.