基于加速度传感器的人体姿态识别系统

Human Body Attitude Identification System Based on Accelerometer

设计了一种基于加速度传感器的人体姿态识别系统,系统采用MPU-6050九轴运动处理传感器采集人体的运动轨迹信号。当人体在做日常的一般运动以及突发跌倒时,三轴陀螺仪和三轴加速度计分别将其测量的模拟量转化为3个16位可输出的数字量,得到的结果直接送入单片机进行相关处理,然后经nRF2401无线传输模块传送到接收计算机。对人体的各种日常姿态进行检测,通过对各项实验数据进行观察分析,确定了人体跌倒的阈值以及根据人的生理特征设置了发生跌倒的倾斜角度,两者相结合进行人体姿态的识别与判读。在试验条件下,系统对跌倒识别的准确率平均为95%。

A kind of human body gesture recognition system is designed based on acceleration sensor,in which MPU-6050is used for collecting the nine axis motion signal of human body.When human conducts daily movement or sudden fall,the three 16-bit digital signal that translated from analog signal will be output,which is measured by three-axis gyroscope and three-axis accelerometer,and MCU processes the three 16-bit digital signal that is transmitted from the MPU-6050,the results are directly transmitted to the receiving computer by nRF2401.The threshold value and angle of inclination of human falling are confirmed by detecting all kinds of daily gestures of the human and analyzing the experimental data.Identification and interpretation of the human gesture are carried out by combining the threshold value with the angle of inclination.Under laboratory conditions,the average accuracy of the system is 95%.