论文部分内容阅读
目的:本研究旨在开发和验证一种在训练中能协助快速反馈链球线速度的方法。方法:通过测量链球顶部三维位置数据,计算链球的线速度和锁链作用力。用这些数据建立两种线性回归模型,并利用链锁作用力预测链球速度:一种模型采用通过调控的数据建立模型,使速度和作用力的最大值和最小值一致(移位模型);另一种模型的数据则未经过调控(非移位模型)。用这两种模型并通过直接测量锁链作用力预测链球球速。通过比较预测的和通过链球的位置数据测得的链球速度,评估模型的准确性。收集每人每次投掷球速的预测值与计算值,计算其差值的多元相关系数(CMC)和均方根(RMS)的平均值,以此来评估预测值的准确性。结果:两种模型均呈现出较高的CMC(0.96和0.97)和相对较低的RMS值(1 27 m/s和1.05 m/s)。此外,球速预测值与测量值的平均百分比差异在非移位和移位组中分别为6 6%和4.7%。同时通过两种模型及三维定位数据计算链球出手速度的均方根差值。非位移模型和位移模型中,预测值和计算值的均方根差值分别为0.69 m/s和0.46 m/s。结论:本研究成功推导并验证一种能够从直接测得的链锁作用力预测链球速度的有效方法。开发的两种线性回归模型都能预测速度,但使用移位模型更加准确。
Purpose: This study aims to develop and validate a method that can assist in the rapid feedback of the speed of a hammer ball during training. Methods: By measuring the top three-dimensional position of the top of the hammer ball chain data to calculate the chain speed and chain forces. Using these data, two linear regression models are established and the chain ball force is used to predict the speed of the chain: one model uses the data from the regulation to establish a model that maximizes the maximum and minimum of velocity and force (shift model); and the other The data for one model is not regulated (non-shifted model). Use both models and predict the hammer speed by directly measuring the force of the chain. The accuracy of the model is assessed by comparing the predicted ballistic speed with the position data of the hammer. The predictive value and calculated value of each throwing speed are collected and the multivariate correlation coefficient (CMC) and the average of root mean square (RMS) values of the difference are calculated to evaluate the accuracy of the predicted value. Results: Both models showed higher CMC (0.96 and 0.97) and relatively lower RMS values (127 and 1.05 m / s). In addition, the average percentage difference between predicted and measured ball velocity was 66% and 4.7% respectively in the non-shifted and shifted groups. At the same time, two models and three-dimensional positioning data are used to calculate the root-mean-square difference of hammer speed. In the non-displacement model and the displacement model, the root-mean-square differences between predicted and calculated values are 0.69 m / s and 0.46 m / s, respectively. Conclusion: This study successfully derives and validates an effective method to predict the speed of the chain ball from the directly measured chain force. Both linear regression models developed can predict the speed, but using the shift model is more accurate.