融合光流与MARG传感器的三维姿态测量研究

Research on Three-Dimensional Attitude Measurement Incorporating Optical Flow and MARG Sensors

光流传感器可获取其载体的三维运动信息,近年来被广泛用于无人机等领域,但光流测量值中包含的三维角速度信息未能得到充分利用。磁强计、加速度计与陀螺仪的组合(又称MARG传感器)同样常被用于无人机实现三维姿态测量,但低成本MARG传感器精度有限。为提高姿态测量精度,提出一种采用光流传感器获取三维角速度的测量模型,并根据最小方差准则将光流传感器提取出的角速度与陀螺仪测量值进行融合,进而将融合后的角速度与加速度计、磁强计测量值进行卡尔曼滤波,从而获得三维姿态估计。实验表明,采用光流传感器能有效地获取三维角速度;将光流传感器与陀螺仪融合后可使三轴角速度均方根误差不大于1°/s;将光流传感器与MARG传感器融合后能使三维姿态角均方根误差限制在1°以内,有助于提高姿态测量的冗余度和可靠性。

Optical flow sensor can acquire the information of three-dimensional motions of its carrier,and it has been widely used in unmanned aerial vehicle(UAV)and other fields.However,the information of three-dimensional angular velocity that contained in optical flow measurement has not been properly utilized.On the other hand,the combination of magnetometer,accelerometer,and rate gyro(abbreviated as MARG sensor)is also frequently used in UAV for attitude measurement,but the accuracy of low-cost MARG sensor is limited.To improve the accuracy of attitude measurement,a measurement model utilizing optical flow sensors is proposed,which can acquire three-dimensional angular velocity.The angular velocity acquired from optical flow sensors is fused with the measurement of gyroscope,and then the fused angular velocity is processed by Kalman filter along with the measurement of magnetometer and accelerometer according to the minimum variance principle to get three-dimensional attitude estimation.Experiment results show that the three-dimensional angular velocity can be effective acquired through optical flow sensors.The root mean square errors(RMSE)of three-axis angular velocity can be less than 1°/s by incorporating optical flow sensor with gyroscope,while the RMSE of three-dimensional attitude can be limited within 1°by incorporating optical and MARG sensors,which can help to enhance the redundancy and reliability of attitude measurement.