融合体能评估与强化学习的个性化篮球教学训练策略

打开文本图片集

【中图分类号】G841 【文献标识码】A 【文章编号】2095-2813（2025）19-0018-05

[Abstract] With the increasing personalization of sports training demands，designing precise training programs based onathletes'diferent physicalcharacteristics hasbecomeacrucial topicinmodersports training.Basketballplayers' physical fitness levels directlyafects theeffectivenessoftheir technicalandtactical execution.Therefore，personalized training optimization is crucial for improving competitive performance.This paper proposes a personalized basketball training optimization framework based on Reinforcement Learning （RL），using physical fitnessasssment data to guide training strategyselection.Toovercomethe problem of insuficient training data，amulti-level data augmentation strategy is adopted，incuding SMOTEalgorithm，GANnetworks，and Gaussiannoiseijection techniques，toepand original samples 5-10 times. This research designs an automated training decision system based on Q-learning algorithm， which can adjust the training plan inreal time according tothe physicalfitness status of athletes.Experimental results show that the experimental group's maximum oxygen uptake has significantly increased compared to before the experiment （P<0.001） ， while mixed sample training models outperformed pure real sample models by approximately 30% This research provides new ideas and practical references for inteligentandpersonalized development in basketball training.

[Keywords]Basketball trainng;Personalizedtraining;Reinforcementlearing;itnessassessment;Dataenhancement

作者简介：陈建桦（1997—），男，本科，研究方向为运动训练学、体育数据分析与技术应用。（剩余8049字）