摘要
智能手机内置的多种传感器可测量运动、屏幕方向和各种环境条件,被广泛应用于智能感知中对人体行为的识别。但在实际应用中采集的运动传感器数据与设备的放置位置和方向相关,通过对欧拉角到旋转矩阵的转换方法进行研究,给出使用旋转角将加速度传感器数据转换到统一参考坐标系的方法,减少设备方位对采集数据的影响;并结合人体基本行为数据集的识别过程验证转换方法的有效性,对分类识别准确率有明显提高。
Smart phones are equipped with a rich set of sensors,which could measure movement、directions and environmental conditions.These sen⁃sors can be used for human activity recognition in the intelligence perception domain.Based on the rotation matrix(Euler Angle Conver⁃sion)derived from the orientation angles of gyroscopes and orientation sensors,we convert raw accelerometer data from the mobile phone coordinate system into the uniform reference coordinate system.In this paper,we specifically focus on human activity recognition on mo⁃bile phones independent of phones orientations,and describes how more accurate results of basic activity recognition can be achieved with transformed accelerometer data.
作者
衡霞
李宏斌
HENG Xia;LI Hong-bing(School of Computer Science and Technology,Xi'an University of Posts and Telecommunications,Xi'an 710121;Shaanxi Key Laboratory of Network Data Analysis and Intelligent Processing,Xi'an University of Posts and Telecommunications,Xi'an 710121;Basic Medical Department,Shaanxi University of Chinese Medicine,Xianyang 712046)
出处
《现代计算机》
2021年第5期3-7,共5页
Modern Computer
基金
国家自然科学基金项目(No.61373116)
陕西省教育厅专项科研计划项目(No.16JK1706)
陕西省咸阳市重大科技创新专项(No.2017K01-25-3)
横向项目:基于多传感器数据融合的智能家居安防系统。
关键词
加速度传感器
信息转换
行为识别
支持向量机
Acceleration Sensor
Information Transformation
Activity Recognition
Support Vector Machines