Skip to main navigation Skip to search Skip to main content

TaekAI: Revolutionising Taekwondo Training with AI

  • V. Arjun Doss
  • , Amulya Doss
  • , H. Shloka Gowda
  • , Tarun Kumar

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

Abstract

Taekwondo, a traditional Korean martial art, emphasises powerful kicks and punches, requiring mastery in three core areas: Poomsae (technique patterns), kyorugi (sparring), and gyeokpa (breaking). This paper introduces TaekAI, an innovative application that blends Taekwondo training with PoseNet, an AI-powered pose detection model. TaekAI uses PoseNet to detect 17 key body joints, creating a precise skeleton picture of the user's position and calculating essential angles to assess alignment with certain Poomsae movements. The system consists of many modules—Detect, Determine Joints, Calculate Angles, Correct, and Repeat—that deliver valuable input to practitioners at all levels. TaekAI facilitates an ongoing feedback loop that helps users enhance their Taekwondo skills through notifications and corrections. TaekAI revolutionises poomsae practice by analysing the users’ form, providing real-time feedback, and suggesting posture corrections. The app evaluates poomsae accuracy, timing, and technique performance, giving valuable insights to improve practice sessions. The paper also explores TaekAI’s technical framework, system design, and potential applications in enhancing Taekwondo practice.

Original languageEnglish
Title of host publicationStudies in Systems, Decision and Control
PublisherSpringer Science and Business Media Deutschland GmbH
Pages1075-1087
Number of pages13
DOIs
Publication statusPublished - 2025

Publication series

NameStudies in Systems, Decision and Control
Volume291
ISSN (Print)2198-4182
ISSN (Electronic)2198-4190

Keywords

  • Human pose estimation
  • Martial arts
  • Poomsae
  • PoseNet
  • Taekwondo

Fingerprint

Dive into the research topics of 'TaekAI: Revolutionising Taekwondo Training with AI'. Together they form a unique fingerprint.

Cite this