手机内置加速度传感器数据的空间坐标转换算法

发布时间：2018-08-20 15:49

【摘要】：手机内置加速度传感器坐标系固定于设备自身,其采集的数据因手机姿态的改变而不断发生漂移,受此影响即使同一运动过程,加速度数据也难以同前一个时刻保持一致。为解决该问题,提出利用空间坐标转换算法将加速度数据从手机坐标系映射至惯性坐标系,从而确保数据在手机任意姿态下均能准确反映实际的运动状态。为验证该方法的有效性,设计一种手机传感器数据在线采集与实时处理新方法,实现Matlab中数据动态特征的实时观测及算法性能的在线评估。利用此方法,在旋转实验中分别测试方向余弦与四元数两种算法的可行性,并在计步器实验中进一步测试四元数算法性能。实验结果表明,基于方向传感器数据的方向余弦算法因测量范围限制,不能实现全方位空间坐标转换;而基于旋转矢量传感器数据的四元数算法则能够实现全方位转换,且转换后的加速度对步态识别率达到95%以上,较准确地反映了实际运动状态。
[Abstract]:The built-in acceleration sensor coordinate system is fixed on the device itself, and the data collected by the mobile phone constantly drift because of the change of the mobile phone attitude. Even if the acceleration data is in the same motion process, it is difficult to keep the acceleration data consistent with the previous time. In order to solve this problem, the acceleration data is mapped from the mobile phone coordinate system to the inertial coordinate system by using the spatial coordinate transformation algorithm, so as to ensure that the data can accurately reflect the actual motion state of the mobile phone in any attitude. In order to verify the effectiveness of this method, a new on-line data acquisition and real-time processing method for mobile phone sensor is designed to realize real-time observation of data dynamic characteristics and on-line evaluation of algorithm performance in Matlab. Using this method, the feasibility of direction cosine algorithm and quaternion algorithm is tested in rotation experiment, and the performance of quaternion algorithm is further tested in the pedometer experiment. The experimental results show that the directional cosine algorithm based on the direction sensor data can not realize the omnidirectional spatial coordinate transformation due to the limit of the measurement range, while the quaternion algorithm based on the rotation vector sensor data can realize the omnidirectional transformation. The recognition rate of gait is more than 95%, which reflects the actual state of motion more accurately.
【作者单位】：兰州理工大学计算机与通信学院;
【基金】：国家自然科学基金资助项目(61262016) 甘肃省自然科学基金资助项目(1208RJZA239)~~
【分类号】：TP212.9

【相似文献】