A new artificial intelligence framework designed for physical education (PE) integrates wearable biosensing with a multi-objective Actor-Critic decision model to provide personalized training and injury prevention. By reconciling heterogeneous data sampling rates—ranging from 1 Hz to 100 Hz—the system addresses the complexities of classroom-scale environments, including physiological variation and latency constraints. The integrated framework focuses on individual baseline calibration to balance training effectiveness with safety measures.
AI Framework Improves Fitness Outcomes and Reduces Injuries
In an eight-week controlled teaching study involving 120 participants, the implementation of this AI-driven framework yielded significant differences between the intervention and control groups. Students utilizing the system demonstrated higher improvement efficiency in fitness outcomes, with a recorded margin of 32.4%. The framework showed a marked impact on student safety, as injury incidence in the intervention group was 1.7%, compared to 13.3% in the control group.
Technical performance metrics indicate the system is capable of managing high-density environments. On a self-built campus dataset, the framework achieved a risk-warning recall of 95.8% and plan adaptability of 94.7%. During operation, the system supported 200 concurrent users while maintaining a mean response time of 185 milliseconds.
Single Institution Design Limits Generalizability of Results
While the study demonstrates the feasibility of this technology within a university PE setting, researchers noted that the findings are constrained by the single-institution design. The reliance on this specific environment limits the broader generalizability of the results. The study highlighted that teacher awareness of allocation during the intervention introduces variables that complicate causal interpretation.
The framework distinguishes itself as an integrated design tailored specifically for physical education rather than a new learning paradigm. Future applications of such technologies in sports and health contexts remain subject to the ongoing challenges of sensor reliability and the necessity of human oversight to ensure that AI-generated prescriptions remain valid and interpretable.
