At present, there are problems in competitive aerobics training, such as unclear basis for training decisions, delayed feedback, and insufficient personalization. To this end, this paper designs and implements a training decision-making and evaluation system that integrates artificial intelligence (AI) and intelligent Internet of Things (IIoT) technologies. The system adopts a four-layer architecture: the perception layer collects motion and physiological data through multiple sensors such as IMU, electromyography, and heart rate belt; the network layer uses 5G and Wi-Fi 6 transmission and edge preprocessing; the platform layer combines convolutional neural network (CNN) and long short-term memory (LSTM) network for action recognition and training status evaluation; the application layer generates personalized training plans based on reinforcement learning algorithms and provides visual decision support. In the experimental comparison, the training effect of athletes is significantly improved after the new system is adopted: the body control score increases from 78 to 91; the training difficulty score increases from 81 to 96; the performance style score increases from 75 to 86; the training effect increases by an average of 16.63%. In addition, the survey shows that the athletes' satisfaction with the new system exceeds 83% in each group, with the highest reaching 90%. Research has shown that the training decision-making system combining AI and IIoT has significant value in improving the scientificity and effectiveness of competitive aerobics training.
Haiyan Li (Tue,) studied this question.