在人体运动测量中应用四元数扩展卡尔曼滤波

Application of quaternion extended kalman filtering in human sport measurement

提出了一种适合多运动传感器方位测定系统的四元数扩展卡尔曼滤波,该方位测定系统适合人体运动领域的研究,它是基于iNEMO整合套件,集成了三轴陀螺仪、三轴加速度计和三轴磁强计。在该文的扩展卡尔曼滤波过程中,通过加速度计、磁力计的量测噪声进行陀螺仪角度修正,使测量角度值逐渐逼近真实角度大小,同时量测噪音的协方差矩阵设计,可以使方位测定系统辨别静止和运动,增加了可靠性。结合人体前臂屈伸运动的测试,验证了算法的有效性,加之其小尺寸、多功能、低功耗的特点,证明基于MEMS的运动传感器在人体运动测量领域的应用前景良好。

A quaternion extended Kalman filtering algorithm suitable for multiple motion sensor orientation detection system is proposed.The orientation system adapts to the research of human motion.The hardware is based on iNEMO sensors fusion suite,which integrated 3-axis gyroscope,triaxial accelerometer and magnetometer.In the procedure of extended Kalman filter,the values of angle will approach the true value gradually through modifying angle of gyroscope by the measuring noise of accelerometer and magnetometer,at the same time,the design of covariance matrix on measuring noise can help the orientation system distinguish the static and movement state.With the test of movement of flexion and extension of forearm,this algorithm was verified effective.With the merits of small size,multi-function and low loss,the motion sensors based on MEMS are proved to have a favorable future in the field of human motion measurement.