基于四元数互补滤波的无人机姿态解算

Attitude estimation based on quaternion complementary filter for UAV

针对无人机低成本姿态解算这一基本问题,考虑到传统姿态算法运算量大、不易调试,采用微惯性单元(MEMS)测量无人机原始姿态数据,采用基于四元数的互补滤波算法,有效降低姿态解算的运算量,实现MEMS各传感器的信息融合。从理论上证明了基于四元数的互补滤波器的稳定性,分析了滤波器的性能。采用无人机真实数据验证了算法的有效性,解算得到的俯仰角、滚转角精度小于1°,航向角精度小于2°。与传统姿态算法比较,本算法简单有效、运算量小、易于调试。

Aiming at the basic problem of attitude estimation of unmanned aerial vehicles, in consideration of the quality of the traditional algorithm which is complex and hard to debug, MEMS is used to measure the data of UAV primitive attitude. Quaternion complementary filtering algorithm is applied to reduce the operation quantity of attitude estimation and make the fuse of information collected by every sensor. The complementary filter stability based on quatemion is theoretically proved. And the quality of filter was analysed here. The effectiveness of the algorithm used here is testified by real data from UAV. The precision of pitch and yaw are less than 1° and 2° respectively. Compared with traditional attitude algorithm, this algorithm is simple and effective, small computational complexity, and easy to debug.