一种新的羽毛球拍挥拍轨迹识别方法

发布时间：2018-01-12 15:25

本文关键词：一种新的羽毛球拍挥拍轨迹识别方法　出处：《东北大学学报(自然科学版)》2017年01期 　论文类型：期刊论文

【摘要】：为了更加精确地识别羽毛球拍的运动轨迹,提出一种多传感器融合的挥拍轨迹识别方法(TRM,trajectory recognition method).使用放置在球拍拍柄底部的智能设备进行数据采集,该设备由加速度计、陀螺仪和磁力计的多传感器组成.首先,使用加速度计和磁力计对陀螺仪进行修正,改进基于卡尔曼滤波器的四元数姿态解算精度并得到姿态角.其次,使用旋转矩阵去除加速度的重力分量,再对加速度数据进行频域数值积分得到速度和位移,采用最小二乘拟合多项式剔除数值积分过程中产生的累积误差.最后,结合姿态角、速度和位移,识别出挥拍轨迹.实验结果表明,提出的TRM具有良好的有效性.与传统时域积分轨迹识别方法相比,TRM的挥拍轨迹识别率更高.TRM可以更准确地进行羽毛球拍运动轨迹识别.
[Abstract]:In order to identify the motion trajectory of badminton racquet more accurately, a multi-sensor fusion method named TRM is proposed. Trajectory recognition method uses a smart device placed at the bottom of the racket handle to collect data from an accelerometer. Firstly, gyroscopes are modified by accelerometers and magnetometers to improve the precision of quaternion attitude resolution based on Kalman filter and get attitude angles. The rotation matrix is used to remove the gravity component of the acceleration, and the velocity and displacement are obtained by the numerical integration of the acceleration data in the frequency domain. The least square fitting polynomial is used to eliminate the accumulated errors in the process of numerical integration. Finally, the swing trajectory is identified by combining attitude angle, velocity and displacement. Compared with the traditional time-domain integral trajectory recognition method, the proposed TRM has a higher recognition rate of swing trajectory. It can be used to identify the motion trajectory of badminton racquet more accurately.
【作者单位】：东北大学计算机科学与工程学院;沈阳大学信息工程学院;
【基金】：国家科技支撑计划项目(2012BAH82F00) 辽宁省科学技术计划项目(2015401039) 沈阳市科技专项(F15-199-1-03)
【分类号】：TP391.41;TP212.9
【正文快照】： 羽毛球运动是一项技术主导型体育竞技运动,具有对抗性强、运动姿态灵活度高等特点.挥拍动作具有灵活性高、重复性弱的特点,从而导致运动员无法客观衡量自己每次挥拍动作的技术特点,只能靠大规模重复动作,自己领悟技术要领.现有羽毛球挥拍轨迹识别主要采用计算机视觉技术.比如,

【相似文献】