Time Synchronization Algorithm for the Skiing Monitoring System
Yun Yang,Rui Cheng,Jian He,Chuncheng Li,Xiaojun Qiao,Xiaojuan Hou,Jijun Xiong,Xiujian Chou
DOI: https://doi.org/10.1109/tim.2022.3173635
IF: 5.6
2022-05-27
IEEE Transactions on Instrumentation and Measurement
Abstract:A skiing monitoring system composed of multiple sensors can feedback videos, kinematic, and physiological parameters synchronously. Also, it is a significant technology for improving the skills of skiers. The time synchronization between diverse sensors is a prerequisite for comprehensive analysis and feedback. The global navigation satellite system (GNSS) and the inertial measurement unit (IMU) sensor are the two key components of the skiing monitoring system. However, due to uncertainties in the quasi-random human motion signals of skiing, it is still a challenge to achieve fully automatic synchronization between the GNSS and independent IMU without imposing any additional burden on skiers. In order to provide a solution that can be used in daily ski training, a time synchronization algorithm based on cross correlation analysis and "0/1 speed dichotomy" (CSD) is proposed. For the first time, the quasi-random human motion signals of skiing collected by the GNSS and independent IMU achieve automatic time synchronization without imposing any additional burden to the athletes. In this method, the sampling points of IMU and GNSS are divided into static state interval and motion state interval by a static interval detection algorithm, and then, the speed is assigned a value of 0 or 1 depending on the interval in which the sample point is located. It is no longer necessary to reconstruct the speed by integration operation using the acceleration collected by IMU. The algorithm is verified by using 21 training datasets from ten national athletes. The results show that even in the presence of ski resorts and athletes changing, the average error of CSD is 84.65 ms and the average running time of CSD is only 58.77 ms. The average error is further reduced to 34.29 ms with a standard deviation of 33.64 ms when the sampling rates of GNSS and IMU are increased. Compared to time synchronization algorithms based on high-pass filters, zero velo- ity update (ZUPT), and improved ZUPT, CSD takes less time to run with smaller synchronization errors and standard deviations when test athletes and snowfields change. CSD shows distinct advantages, including low computational cost, good robustness, and high accuracy. Therefore, CSD is adopted for application in a synchronous skiing monitoring system. The system can perform collection, synchronous analysis, feedback of various data, and videos of skiers without affecting the skiers' actions. At present, the system has been used in the daily training of a professional cross-country skiing team and has provided some suggestions for athletes to improve their physical distribution and technical actions.
engineering, electrical & electronic,instruments & instrumentation