Honors Theses

Title

Golf Glove Data-based Swing Classification through Machine Learning

College

James Worth Bagley College of Engineering

Department

Department of Electrical and Computer Engineering

Degree

Bachelor of Science (B.S.)

Major

Computer Engineering

Document Type

Honors Thesis

Abstract

A golf swing is biomechanically complex. Professional swing training is expensive for the average golfer. With the growing development of small inertial sensors and powerful microprocessors with built-in wireless communication protocol support, embedded devices are becoming suitable for tough tasks like motion tracking. The proposed solution consists of a sensor-packed golf glove. To evaluate the efficacy of the proposed solution, a recurrent neural network is developed that uses a learning model to identify golf swings that produce a slice, the most common golf swing error. A motion capture system was used as the professional baseline for the evaluation. Barely falling short of the professional solution’s performance, the proposed solution showed potential to become a portable and economical alternative.

Publication Date

5-1-2019

First Advisor

Ball, John E.

Second Advisor

Burch V, Reuben F.

Third Advisor

Elder, Anastasia

This document is currently not available here.

Share

COinS