基于运动捕捉系统的多旋翼飞行器实时测姿算法

摘要

旋翼飞行器的室内自主飞行是目前研究的热点之一.在室内飞行过程中,飞行器姿态和位置信息可以通过运动捕捉系统(Motion Capture System, MCS)来进行实时测量,从而为机载低成本MEMS惯性导航系统提供校正信息.结合四旋翼飞行器的结构特性,提出了一种五点实时测姿算法.相对于目前MCS常用的测姿算法,该算法可以降低标记点安装引起的测姿误差.室内实验结果表明,该算法测姿精度高,并且能够有效实现四旋翼飞行器室内动态实时姿态测量,具有较好的工程应用价值.%Much attention has been paid to indoor autonomous flight which is one of the rotorcraft research. The real-time information of rotorcraft obtained from MCS when flying indoors contains position and attitude,and plays a role as the calibration information reference for the airborne low cost MEMS inertial navigation system. With respect to the structural characteristics of quadrotor,a five markers novel real-time algorithm proposed in this paper can reduce the attitude error caused by the installation error well,compared with the conventional estimation algorithm in MCS. As demonstrated by the results of indoor experiments,this proposed algorithm not only actively improves the precision accuracy,but also allows ef-ficient indoor real-time estimation of flying vehicles' attitude. Therefore, this proposed algorithm has good potential in ap-plied engineering.