摘要
获得精准的姿态角信息是四旋翼飞行器控制研究的关键。针对传统算法对四旋翼飞行器高速运动过程的姿态解算精度较低的问题,提出一种基于梯度下降的自适应变步长姿态融合算法。该算法可根据运动加速度的大小自适应地调整对加速度计测量数据的信任程度,从而提高姿态解算精度。实验结果表明,在解算姿态航向参考系统实验设备的偏航角时,自适应梯度下降算法的解算精度较传统互补滤波算法和梯度下降算法分别提高了0.557°和0.708°。
Obtaining accurate information of attitude angle is the key to control research of quadrotor aircraft.To solve the problem of low accuracy of attitude calculation in the process of high-speed motion of quadrotor by traditional algorithm,an adaptive variable step-size attitude fusion algorithm based on gradient descent is proposed.The algorithm can adaptively adjust the trust degree of accelerometer data according to the motion acceleration,thus improving the accuracy of attitude calculation.The experimental results indicate that the accuracy of the adaptive gradient descent algorithm are 0.557°and 0.708°higher than that of the traditional complementary filtering algorithm and gradient descent algorithm when calculating the yaw angle of the test equipment of attitude and heading reference system.
作者
高怡
李东航
GAO Yi;LI Donghang(School of Electronic Engineering,Xi’an Shiyou University,Xi’an 710065,China)
出处
《电子设计工程》
2022年第10期8-11,16,共5页
Electronic Design Engineering
基金
国家自然科学基金(51604226)
陕西省自然基金面上项目(2018JM5064)。
