基于AHRS的人体姿态捕捉系统的设计与实现  

Design and implementation of a human posture capture system based on AHRS

作　　者：张庆宾[1] 田国会[1] 张衍儒[1] 

机构地区：[1]山东大学控制科学与工程学院,济南250061

出　　处：《高技术通讯》2013年第12期1292-1298,共7页Chinese High Technology Letters

基　　金：国家自然科学基金(61075092);山东省自然科学基金(ZR2011FM011);山东大学自主创新基金(2011JC017)资助项目

摘　　要：研究了利用穿戴式传感器进行人体姿态捕捉的问题。设计实现了一种被称做姿态航向参考系统(AHRS)的传感器模块,将AHRS模块绑定在人体的主要运动关节,其采集的陀螺仪、加速度计和磁力计信号经过基于小波变换的降噪处理后,对三种传感器进行基于扩展卡尔曼滤波算法的数据融合,解算出人体关节的运动姿态。通过WIFI模块将各个关节的姿态数据发送到计算机,上位机软件利用OPENGL技术建立了人体三维骨骼模型,并根据接收的数据对人体姿态与骨骼模型进行坐标转换和关节解耦,将人体姿态动作直观地显示出来。实验结果表明,所设计的惯性传感器模块AHRS的检测精度高,稳定性好,没有数据漂移和累积误差,所提出的姿态捕捉系统能够准确地检测人体姿态动作,反应速度快,满足实时性的要求。The problem of human posture capture using wearable sensos was studied. The attitude and heading reference system（ AHRS）, a sensor module, was designed and implemented, and some AHRS modules were attached to a user＇ s main joints for obtaining signals of a module＇ s three sensors of gyroscope, accelerometer and magnetometer. After the noise reduction processing based on the wavelet transform, the three kinds of sensor data were fused with the extended Kalman filter algorithm for calculating the movement posture of the user＇ s joints. Each joint posture data was sent to the computer by using the WIFI module, and the OPENGL technology was applied to building of the three-dimensional skeletal model. According to the received data, the coordinate transformation was carried out between human posture and the skeletal model, and then the human posture was displayed visually. The experimental results show that the inertial sensor module AHRS has the high detection precision and stability, with no data drift and cumulative error. By using the proposed method, the human posture can be captured accurately and effectively, meeting the real-time requirement.

关 键 词：姿态航向参考系统(AHRS) 小波变换 数据融合 扩展卡尔曼滤波器 人体姿态捕捉 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]