Real-time Monitoring and Analysis of Track and Field Athletes Based on Edge Computing and Deep Reinforcement Learning Algorithm

TL;DR

An IoT-optimized system that integrates edge computing and deep learning algorithms that achieves efficient motion recognition and real-time feedback is proposed, addressing the limitations of traditional monitoring systems in terms of real-time performance and accuracy.

Abstract

This research focuses on real-time monitoring and analysis of track and field athletes, addressing the limitations of traditional monitoring systems in terms of real-time performance and accuracy. We propose an IoT-optimized system that integrates edge computing and deep learning algorithms. Traditional systems often experience delays and reduced accuracy when handling complex motion data, whereas our method, by incorporating a SAC-optimized deep learning model within the IoT architecture, achieves efficient motion recognition and real-time feedback. Experimental results show that this system significantly outperforms traditional methods in response time, data processing accuracy, and energy efficiency, particularly excelling in complex track and field events. This research not only enhances the precision and efficiency of athlete monitoring but also provides new technical support and application prospects for sports science research.

Figures (7)

• Figure 1: MarineYOLO Network Architecture Diagram.
• Figure 2: Example of wearable device and sensor setup for athlete monitoring system: (a) Components of the wearable device, including the Microcontroller Unit (MCU), heart rate sensor, accelerometer, gyroscope, physiological data sensor, and wireless communication module. (b) Internal structure of the sensor chip, including the NFC chip, Low Dropout Regulator (LDO), amplifier, sensor, and coil. (c) An athlete wearing multiple sensors on their body, connected via Bluetooth to a mobile device, demonstrating the practical application of sensors in real-time data collection.
• Figure 3: The comparison of three deep learning classifiers used in the athlete monitoring system. (a) RF model illustrating the classification of athlete actions through multiple decision trees, followed by ensemble decision making for final action class determination; (b) GB model demonstrating the training phase, data splitting into subsets, and the final model training using these subsets for enhanced classification performance; (c) CNN model representing the process of pose classification, starting from athlete pose image input, followed by convolution and pooling layers, and culminating in the final classification through global average pooling.
• Figure 4: Comparison of Average Response Times for Different Algorithms Across Various Time Periods.
• Figure 5: Speed and Precision Comparison between SAC-Based and Non-SAC Systems across Various Track and Field Events. (a) demonstrates the time efficiency in milliseconds for different events, (b) highlights the precision percentage in detecting and analyzing athletes' performances.