Research On Human Posture Calculation Algorithm Based On Multi-sensor Information Fusion

The measurement and identification of human posture is the process of recovering the position of human joint points,which mainly describes human posture and predicts human behavior,and plays a very important role in various fields such as human-computer interaction,virtual reality,and medical rehabilitation.The technical means adopted are mainly video technology and sensor technology,and with the continuous updating and development of microelectromechanical systems(MEMS)technology,the human posture solution and recognition system composed of magnetic force,acceleration and gyroscope based on MEMS inertial sensor technology has become a research hotspot.In the process of human posture detection,the error of magnetometer and accelerometer is relatively large when measuring the posture angle,with poor dynamic performance and good steady-state characteristics;the dynamic characteristics of gyroscope are good,but the steady-state performance is poor due to the accumulation of integral drift error,in order to achieve the perfect measurement effect of dynamic and static posture,the angle solved by accelerometer and magnetometer will be differed from the angle solved by gyroscope,and the above deviation will be sent to the regulator The above deviation will be sent to the regulator for mediation,and then the data after mediation will be used to correct the gyroscope,so that the angle solved by the gyroscope and the angle solved by the accelerometer will be the same,thus realizing the fusion and complementary function of multiple sensor measurement data.In addition,for the problem that the PID controller cannot adjust the parameters online,the fuzzy control combined with the PID regulator is introduced to form a fuzzy PID controller that can adjust the parameters online adaptively.In this paper,firstly,a set of lower computer unit based on MPU9150 nine-axis inertial sensor module and controller is constructed,and finally constitutes a human motion information acquisition system by communicating with the upper computer through the RF unit.Then,a human posture angle solving algorithm based on fuzzy PID regulation and complementary filtering is designed to solve the motion data into human posture angle data.Finally,MATHLAB software is used to build a simulation experiment platform,which is used to compare the experimental data of fuzzy PID complementary filtering algorithm and common PI filtering algorithm.The experimental results show that the posture angle after processing by fuzzy PID complementary filtering algorithm has good stability and high accuracy.Translated with www.Deep L.com/Translator(free version)...