Research on the Application and Strategies of AI Empowerment in Sanda Technique Teaching and Correction in Higher Education under the Guidance of Educational Reform

DOI: https://doi.org/10.62517/jhet.202515610

Author(s)

Donghui Tan1,*, Zherao Le2

Affiliation(s)

1School of Physical Education and Art, Jiangxi University of Science and Technology, Ganzhou, Jiangxi, China 2School of Science, Jiangxi University of Science and Technology, Ganzhou, Jiangxi, China *Corresponding Author

Abstract

Under the context of "Five-Education Integration" and Educational Digital Transformation, this study addresses the issue of insufficient precision and low efficiency in artificial motion correction in Sanda (Chinese Kickboxing) teaching at higher education institutions. Given the exploratory demands of the early stage of educational reform, which emphasizes "small-scale entry point and easy implementation," this research aims to explore the initial application path of AI motion recognition technology in correcting core Sanda techniques, verify its technical feasibility and practical effectiveness, and provide replicable pilot strategies for the integration of sports teaching technology in the early stage of educational reform. A total of 30 non-physical education undergraduate students from the Sanda elective course at Jiangxi University of Science and Technology were selected as participants. The experimental group and control group (15 students each) were randomly assigned based on "random sampling and baseline level matching." There were no significant differences between the two groups in terms of gender composition, sports foundation, previous technique standardization, or physical fitness (P>0.05), ensuring balanced and comparable baseline data. This study employed experimental and mixed research methods, with an experimental period of 8 class sessions (1 teaching cycle). The experimental group adopted the "AI motion recognition-assisted + manual guidance" correction model, where a lightweight motion recognition tool was used to capture core movements such as the straight punch and roundhouse kick in real-time. After quantifying and comparing the movements, immediate correction suggestions were provided. The control group used the traditional "teacher demonstration - manual error correction" model. Data was collected through motion standardization scoring, error correction duration statistics, teacher-student feedback questionnaires, and semi-structured interviews. Statistical analysis was conducted using SPSS 26.0. After the intervention, the experimental group showed a significantly higher improvement in core technique standardization compared to the control group, with a notable reduction in error correction time, demonstrating the effectiveness of the technological application. Both teachers and students exhibited a high level of acceptance of AI motion recognition technology, with teachers reporting that it effectively reduced the burden of individualized guidance. The study indicates that this technology, through precise quantification and real-time feedback, can break through traditional teaching bottlenecks. It is recommended that, in the early stage of educational reform, core technique correction should be used as an entry point, with a lightweight solution promoted for integration.

Keywords

AI Motion Recognition; Higher Education Sanda Teaching; Motion Correction; Educational Reform; Pilot Application

References

[1] Wang Jianjun, Li Xiaotong, Zhang Liguo. Research on the Reform Path of College Physical Education Curriculum under the Perspective of "Five-Education Integration". China Sport Science, 2021, 41(5):78-85. [2] General Administration of Sport of China. 2023 White Paper on Physical and Mental Health of Chinese College Students. Beijing: People's Sports Publishing House, 2023:45-52. [3] Zhang Bingbing, Zhao Baishan. Sanda Action Recognition and Evaluation Method Based on Skeletal Keypoints. Microprocessors, 2023, 44(6):36-39. [4] Hu Ti, Min Hang. Application Research on the Empowerment of Physical Education Teachers' Professional Development through Generative Artificial Intelligence. Journal of Shenyang Sport University, 2025, 44(1):1-8. [5] Xu Weikang. The Legitimacy Dilemma and Response of Artificial Intelligence in Sports Refereeing. China Sport Science, 2022, 42(11):90-97. [6] He Lei. Research on Signal Acquisition and Recognition Method of Intelligent Protective Gear under Impact. Shenyang University of Technology, 2022. [7] Yang Wanpeng, Li Qing, Lei Ming. Research on Feature-Level Fusion LSTM-CNN Method for Human Action Recognition. Electronic Measurement Technology, 2021, 44(17):173-180. [8] Xu Haibo. Innovative Pathways of Integrating Artificial Intelligence into the Reform of Physical Education Teaching in Colleges and Universities. Journal of Lishui University, 2024, 46(2):123-128. [9] Meng Xiangpu, Li Shuo, Yuan Mingzhe, et al. Human Skeleton-Based Action Recognition: A Review and Prospects. Information and Control, 2025, 54(1):1-27. [10] Li Sen. Research on the Application Strategies of Artificial Intelligence in the Reform of College Physical Education Teaching. Hubei Open Vocational College, 2024, 37(9):167-168. [11] Shen Liang, Zhang Bo, Chen Xiangwei. Practical Implication, Theoretical Interpretation and Implementation Path of Chinese-Characteristic Physical Education Curriculum under the Perspective of "Five-Education Integration". Journal of Shanghai University of Sport, 2024. [12] Zhang Yu. Research on the Application of AI Technology in Physical Education Teaching in Primary and Secondary Schools. Journal of The Chinese Society of Education, 2022(5):89-93. [13] Shao Mingjun, Ding Yi, Zhang Zhanping, et al. Research on the Physical Health Status of College Students. New Sports, 2025(10):98-100. [14] Liu W, You D, Fu Z, et al. Sanda Training Program Based on Multi-Sensor Data Fusion in the Internet of Things Environment. International Transactions on Intelligent Systems, 2023, 47:1-15.