Estimating the Standing Long Jump Length from Smartphone Inertial Sensors through Machine Learning Algorithms

Area of Science:

• Sports Science
• Biomechanical Engineering
• Machine Learning

Background:

• The standing long jump (SLJ) is a key metric for assessing motor skills and athletic performance.
• Accurate and accessible measurement tools are needed for athletes and coaches.
• Current methods can be cumbersome or require specialized equipment.

Purpose of the Study:

• To develop a methodology for easily measuring SLJ length using smartphone inertial measurement units (IMUs).
• To leverage machine learning to predict SLJ performance from sensor data.
• To validate the accuracy and reliability of the proposed smartphone-based system.

Main Methods:

• Recruited 114 trained young participants for instrumented SLJ tasks.
• Extracted biomechanical features from smartphone IMU data.
• Utilized Lasso regression to identify key predictors.
• Trained optimized machine learning architectures, including Gaussian Process Regression.

Main Results:

• Achieved an accurate estimation of SLJ length with a Root Mean Square Error (RMSE) of 0.122 m.
• Demonstrated high reliability with Kendall's Tau (τ) < 0.1.
• Confirmed homoscedastic results, indicating consistent model error across different jump lengths.

Conclusions:

• Smartphone sensors provide a feasible, low-cost solution for automatic and objective SLJ performance assessment.
• This technology enables ecological, real-time performance evaluation for athletes and coaches.
• The developed methodology enhances accessibility to sports performance analytics.