Boxing Punch Detection and Classification Using Motion Tape and Machine Learning

Area of Science:

• Sports Science
• Biomechanics
• Wearable Technology

Background:

• Boxing performance analysis often relies on subjective observation or complex equipment.
• Wearable sensors offer a potential solution for objective, real-time data collection.

Purpose of the Study:

• To classify boxing punch types using machine learning algorithms.
• To evaluate the efficacy of a wearable sensor (Motion Tape) for capturing skin-strain data during boxing movements.

Main Methods:

• A human participant study was conducted involving boxing training.
• Subjects performed jabs and hooks with and without striking a heavy bag.
• Skin-strain time history data was collected using Motion Tape and processed by time series classification algorithms.

Main Results:

• Machine learning models successfully classified different punch types based on skin-strain measurements.
• The Motion Tape system demonstrated effectiveness in differentiating between various punches and conditions.

Conclusions:

• Wearable Motion Tape sensors combined with machine learning provide an effective method for classifying boxing punches.
• This system shows significant potential for objective human performance analysis in sports and biomechanics.