Validity of a Commercially Available Inertial Measurement Unit for Artificial Intelligence-Based Trick Detection and...

Area of Science:

• Biomechanics
• Sports Technology
• Skateboarding Performance Analysis

Background:

• Inertial measurement units (IMUs) offer potential for skateboarding performance diagnostics.
• Systematic validation of IMUs in skateboarding is limited.
• The Spinnax Freak IMU system's accuracy needs evaluation for skateboarding applications.

Purpose of the Study:

• To validate the Spinnax Freak IMU system for skateboarding.
• To assess its accuracy in trick detection, classification, distance, speed, height, and airtime.
• To compare IMU data with established reference systems.

Main Methods:

• 23 skateboarders participated in the study.
• Validation involved comparing Spinnax Freak data with laser ranging, 2D video analysis, light barriers, and fixed references.
• Statistical analyses included MAE, RMSE, MAPE, t-tests, Bland-Altman plots, regression, and ICC(3,1).

Main Results:

• The Spinnax Freak system demonstrated high validity for trick event detection and distance measurement.
• Statistically equivalent distance measurements were found compared to the reference.
• Substantial errors were observed in trick classification, maximal horizontal speed, maximal vertical height, and airtime.

Conclusions:

• The Spinnax Freak IMU system is reliable for detecting trick events and measuring distance in skateboarding.
• Current outputs for trick classification, speed, height, and airtime are not suitable for biomechanical interpretation.
• Future developments should focus on algorithmic refinement and improved temporal resolution for enhanced accuracy.