Research On Human Lower Limb Motion Capture Algorithm Based On Inertial Measurements Units

Inertial human motion capture system is an autonomous navigation motion capture system.It does not require an external launch source.It can perform motion capture indoors and outdoors.It is easy to wear,with no site restrictions and no delay.This study is based on the inertial measurements units(IMU)of the human lower limb movement capture algorithm,the purpose is to prepare for the development of the human body lower limb movement capture system,this system is mainly used in postoperative rehabilitation of patients with knee joint injury.This topic has studied related algorithms,improved some algorithms,and designed some new algorithms.Three simulation experiments and two physical experiments have been designed to verify the established algorithm.The main research contents are as follows:First,the calibration algorithm between the IMU and the limb was studied.Researched an existing algorithm for joint position estimation using the least square method.This topic expands the algorithm for estimating the position of spherical secondary joints.In this paper,a new joint axis position estimation algorithm is designed for the problem that the hinge joint axis estimation algorithm is unstable and has poor accuracy.This paper studies the existing problems of the existing IMU initial attitude calibration methods,and designs a new relative IMU relative attitude calibration algorithm for the problems.In this paper,simulation experiments are designed to verify the calibration algorithm.Then,the pose estimation algorithm for a single IMU is studied.This article studies how to build a Kalman filter(KF)model to form a KF algorithm that fuses gyro,accelerometer,and magnetometer signals.For the mathematical model of IMU,this paper designs a linear KF model and a nonlinear KF model.The nonlinear KF model is implemented by using extended Kalman filter(EKF)and unscented Kalman filter(UKF)respectively.In this paper,an adaptive parameter adjuster is designed to form three algorithms: AKF,AEKF and AUKF.In this paper,the simulation experiment is designed to verify and select the algorithm,and the physical experiment is designed to further verify the selected algorithm.Finally,the establishment of human body limb coordinate system and joint angle estimation algorithm are studied.This paper studies the algorithm of establishing the limb coordinate system using the parameters obtained by the calibration algorithm.Based on the the limb coordinate system,This paper also uses the posture estimation information of each IMU to establish the algorithm for solving the lower limb joint angle of the human body.In this paper,simulation experiments are designed to verify the effects of system position calibration error,system attitude calibration error and IMU pose estimation error on the final joint angle estimation error.The joint angle estimation results for motion capture using the IMU and the established algorithm selected in this study are predicted.This article also designed a physical experiment for the verification of the knee joint angle,and verified the algorithm partially.