Affordable Access

Access to the full text

IART: Inertial Assistant Referee and Trainer for Race Walking †

Authors
  • Caporaso, Teodorico
  • Grazioso, Stanislao
Type
Published Article
Journal
Sensors
Publisher
MDPI AG
Publication Date
Jan 31, 2020
Volume
20
Issue
3
Identifiers
DOI: 10.3390/s20030783
PMID: 32023976
PMCID: PMC7038531
Source
PubMed Central
Keywords
License
Green

Abstract

This paper presents IART, a novel inertial wearable system for automatic detection of infringements and analysis of sports performance in race walking. IART algorithms are developed from raw inertial measurements collected by a single sensor located at the bottom of the vertebral column (L5–S1). Two novel parameters are developed to estimate infringements: loss of ground contact time and loss of ground contact step classification; three classic parameters are indeed used to estimate performance: step length ratio, step cadence, and smoothness. From these parameters, five biomechanical indices customized for elite athletes are derived. The experimental protocol consists of four repetitions of a straight path of 300 m on a long-paved road, performed by nine elite athletes. Over a total of 1620 steps (54 sequences of 30 steps each), the average accuracy of correct detection of loss of ground contact events is equal to 99%, whereas the correct classification of the infringement is equal to 87% for each step sequence, with a 92% of acceptable classifications. A great emphasis is dedicated on the user-centered development of IART: an intuitive radar chart representation is indeed developed to provide practical usability and interpretation of IART indices from the athletes, coaches, and referees perspectives. The results of IART, in terms of accuracy of its indices and usability from end-users, are encouraging for its usage as tool to support athletes and coaches in training and referees in real competitions.

Report this publication

Statistics

Seen <100 times