计算机科学
人工智能
运动(音乐)
蓝图
考试(生物学)
表(数据库)
机器学习
粗大运动技能
阿迪模型
姿势
运动(物理)
人机交互
精神运动学习
运动技能
职位(财务)
专家系统
数字化学习
时限
数字媒体
运动评估
模拟
作者
Andy Supriady,Nancy Trisari Schiff
出处
期刊:Sportis
[University of A Coruña]
日期:2026-10-01
卷期号:12 (6): 1-21
摘要
Manual gross motor assessments in schools often consume a lot of teachers' time and energy, so a practical automatic movement evaluation system is needed. This research developed and tested the initial design of NARA-MOVE, a learning media concept that transforms the observation criteria of the Test of Gross Motor Development-2 (TGMD-2) into 2D movement counting rules based on body position detection (pose estimation). Through the first three stages of the ADDIE model, we gathered needs from 12 teachers and 4 principals, developed a movement reference table and an initial blueprint (Version 1.0), and tested it with four experts in physical education, biomechanics, media technology, and digital learning. This design received a feasibility score of 86.9% (SD=0.47) with a high level of expert agreement (Kendall's W=0.86, p<0.01), consisting of 90.0% for learning suitability, 87.5% for movement safety, 87.5% for the AI system, and 82.5% for ease of use. Based on feedback from experts, the design was refined into Version 1.1 by adding a 3-frame motion stabilizer (smoothing filter), an angle tolerance limit of ± 5° to prevent children from becoming easily discouraged during assessments, and local memory processing (RAM) without storing video recordings to maintain student privacy. These validation results prove that the learning logic, movement rules, and privacy protection in NARA-MOVE are accurate and ready to proceed to the application program development stage and direct testing in schools.
科研通智能强力驱动
Strongly Powered by AbleSci AI