低成本无人机姿态解算研究

Low Cost UAV Attitude Algorithm Research

针对小型四旋翼无人机姿态解算数据精度低、缺少余度控制、易发散等问题,提出一种基于GPS、三轴陀螺仪加速度计、三轴磁力计的随机加权自适应滤波算法估计无人机姿态;建立四旋翼无人机姿态旋转矩阵,搭建加速度计和磁力计获取无人机姿态信息的模型,以及采用四元数解算法的陀螺仪定姿解算模型;采用随机加权自适应估计法,依据多元函数求极值定理,在保证总体均方差最小的情况下导出最优随机加权因子,进而解算出姿态角信息,提高四旋翼无人机姿态解算滤波精度与稳定性;仿真与试验结果表明:随机加权自适应滤波与平均值滤波算法相比解算精度更高,输出结果更平稳,且无人机各项预期功能均能正常实现,能够满足四旋翼无人机自主飞行的需要。

In order to improve the drawbacks of low accuracy,lack of redundancy and divergence that occurred in the attitude calculation of micro quadrotor unmanned aerial vehicle （UAV）,a random weighted adaptive filter algorithm is proposed based on GPS,3-axis gyroscope-accelerometer and 3-axis magnetometer,to estimate the UAV attitude.A Four rotor UAV attitude rotation matrix is established,A model for obtaining the attitude information of UAV with accelerometer and magnetometer,and Attitude calculation model of gyroscope sensor u-sing four element method.By randomly weighted adaptive estimation method,based on the multiple function extremum theorem,in ensuring the overall variance under the condition of minimum to derive the optimal random weighting factor,and then calculate the attitude informa-tion,improve the quadrotor UAV attitude estimation filtering accuracy and stability.The simulation and experimental results show that the random weighting adaptive filtering and the average filtering algorithm are more accurate and more stable,and the expected function of the UAV can be realized,which can meet the needs of the autonomous flight of the quadrotor UAV.