基于自适应平方根UKF的微机械传感器融合航姿估计

Low-cost vehicular attitude measurement system using MEMS sensor fusion based on adaptive UKF

提出一种新的基于自适应平方根UKF的微机械传感器组合姿态测量系统.该系统采用3轴微机械陀螺积分得到姿态角,采用3轴微机械加速度计测量重力矢量得到俯仰角和横滚角,分别校正俯仰漂移和横滚陀螺漂移;采用磁强计得到航向角,并与陀螺积分角度融合校正航向陀螺漂移.跑车实验结果表明,基于自适应平方根UKF算法可实时估计机动加速度干扰,并在融合滤波器中进行补偿,能够有效去除车辆机动加速度干扰,姿态角估计精度在±0.6°以内.

To solve the attitude estimation problems for land vehicle use,we proposed a new low-cost attitude estimation system based on the adaptive unscented Kalman filter （UKF）. The system is composed of three MEMS gyroscopes,three accelerometers and magnetometers. The angles were derived by the integration of gyros and the gyro biases were calibrated by accelerometers and magnetometers. To remove the disturbance caused by maneuver accelerations,we proposed a new robust adaptive algorithm to estimate the accelerations in real-time and compensate the disturbance in the accelerometers due to the maneuvering. Experiment results show that the proposed adaptive UKF effectively estimate the attitude and insulate the influence caused by the maneuvering accelerations,and the three angle errors are within ±0.6°.