基于惯性传感器的上肢位置跟踪

Hand Position Tracking Based on Inertial Sensors

中文关键词:  捷联惯性导航  零速检测  卡尔曼滤波  固定区间平滑

英文关键词:SINS  zero velocity detection  Kalman filtering  fixed-interval smoothing

中文摘要:

      为辅助中风病人的康复训练，设计了一种手腕固定惯性测量单元的上肢位置跟踪方案。导航定位算法采用传统的捷联惯性导航解算算法，并在此基础上根据标准康复训练中手臂做屈伸运动时速度周期性为零的特征，引入零速修正技术（ZUPT）。采用姿态检测最优算法检测零速区间，将此时惯性导航解算的速度作为量测值进行Kalman滤波，对系统的速度、姿态、位置、加速度计和陀螺仪常值零偏误差进行估计，将估计结果反馈以修正捷联惯性导航的累积误差。同时在Kalman 滤波的基础上设计了固定区间RTS平滑算法，解决了零速修正引起的运动轨迹的突变问题。实验结果证明，该方案可以有效地实现上肢位置跟踪，在运动时间为108 s的情况下，定位误差为运动路程的0.089%。

英文摘要:

      In order to assist the rehabilitation training for stroke patients,the upper limb position tracking scheme with a wrist fixed inertial measurement unit was designed.Based on the traditional strap-down inertial navigation algorithm,the zero velocity update technology (ZUPT) was also introduced in the navigation and positioning algorithm according to the periodic zero velocity of the arm during the flexion and extension movement in standard rehabilitation training.The stance hypothesis optimal detector algorithm was used for detecting zero velocity interval.The Kalman filter was used to measure the speed of the inertial navigation at this time for the estimation of the the velocity,attitude,position,accelerometer and gyroscope constant deviation of the gyroscope.Then the estimation results were submitted to correct the accumulated error of strapdown inertial navigation system(SINS).At the same time,the fixed interval RTS smoothing algorithm was designed on the basis of Kalman filtering,which solved the mutation of the motion trajectory caused by zero velocity correction.The experimental results show that the proposed scheme can effectively track the upper limb position and the positioning error is 0.089% of the movement distance when the motion time is 108 s.