摘要
提高步频检测的识别率和步长估计的精度是行走轨迹推算定位技术的两个关键。基于手机加速计和行人位置推算原理,利用巴特沃斯滤波器提取手机加速度信号中与行人步态事件相关的信息;借助于峰值探测法识别步频,进而利用步频探测获得的步频点分割加速度信号;通过分析行人行走时的特征,建立了基于加速度波形面积的步长估计模型,以此估计行人的行走路程。实验表明:手机在不同携带位置下,利用频域滤波器提取的信息探测步频的平均识别率达99%以上;在不同步速下,基于面积估计模型的路程估计的平均可靠度在98%以上。
Improving the recognition rate of step frequency detection and the accuracy of step length estimation are two key techniques in the pedestrian dead reckoning and positioning. Based on mobile phone accelerometer and pedestrian position calculation principle, the Butterworth filter is utilized to extract the acceleration signals which are related to pedestrian gait events. The acceleration signal is segmented by the cadence points which are obtained by the peak detection method. With analyzing the characteristics during pedestrian walking, a step-length estimation model based on the area of the acceleration waveform is proposed to calculate the walking length of pedestrian trajectory. Experiment results prove that the average recognition rate in detecting step frequency by frequency domain filter is more than 99%, and the average reliability of the trajectory length estimation based on the area estimation model is above 98% under the condition of asynchronous speed, which could provide reference for high-precision indoor positioning.
作者
郭英
叶瑾
郭金运
王胜利
冯茗扬
GUO Ying;YE Jin;GUO Jingyun;WANG Shengli;FENG Mingyang(College of Geomatics,Shandong University of Science and Technology,Qingdao 266590,China;Ocean Science and Engineering College,Shandong University of Science and Technology,Qingdao 266590,China)
出处
《中国惯性技术学报》
EI
CSCD
北大核心
2019年第6期707-712,共6页
Journal of Chinese Inertial Technology
基金
山东省重点研发计划项目(2018GGX106003)
中国科学院大学生创新实践训练计划(20184001169)
