自适应误差四元数无迹卡尔曼滤波四旋翼飞行器姿态解算方法  

Adaptive Error Quaternion Unscented Kalman Filter Algorithm for Quadrotor Attitude Calculation

作　　者：崔培林 周翟和[1] 吕品[1] 胡斌 CUI Peilin;ZHOU Zhaihe;LU Pin;HU Bin(College of Automation, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China)

机构地区：[1]南京航空航天大学自动化学院,南京211106

出　　处：《西安交通大学学报》2019年第3期97-102,110,共7页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(61703207);航空科学基金资助项目(2017ZC52017)

摘　　要：针对小型四旋翼飞行器在姿态解算时,使用的惯性器件在飞行器起飞、下降等过程中容易受到非重力加速度影响的问题,提出了一种自适应误差四元数无迹卡尔曼滤波的飞行器姿态解算方法。该方法在传统无迹卡尔曼滤波算法的基础上,结合陀螺仪漂移误差模型,使用误差四元数的方法构建无迹卡尔曼滤波的状态方程,选用加速度计和磁强计的输出作观测量,利用其构建无迹卡尔曼滤波的量测方程,并通过自适应调整量测噪声协方差矩阵,减小非重力加速度对姿态解算的影响。实验表明,当飞行器存在非重力加速度时,该方法可以有效减小非重力加速度影对姿态解算的影响。In view of the attitude calculation of a small quadrotor, the inertial device is easily affected by the acceleration of nongravity during takingoff and landing. A new attitude calculation algorithm of quadrotor with adaptive error quaternion unscented Kalman filter is proposed. Following the traditional unscented Kalman filter algorithm, this algorithm combines the gyroscope drift error model and chooses the error quaternion to construct the state equation of the unscented Kalman filter. The measurement equation of the unscented Kalman filter is constructed by using the output signal of accelerometer and magnetometer. Adjusting the noise covariance matrix adaptively, the influence of non-gravity acceleration on attitude calculation is reduced. Experiment shows that when the quadrotor has non gravity acceleration, the algorithm can effectively reduce the influence of non-gravity acceleration shadow on the attitude calculation.

关 键 词：四旋翼飞行器 非重力加速度 误差四元数 无迹卡尔曼滤波 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]