简化超球体SRUKF的GPS/MIMU组合姿态估计

针对加速度计、磁强计易受机动加速度和周围环境影响的问题,提出了一种单基线GPS/MIMU组合姿态估计方法。该方法利用单基线GPS天线代替磁强计提供航向信息,并利用GPS的速度信息对加速度计输出的机动加速度分量进行补偿,构成单基线GPS/MIMU组合姿态确定单元。同时考虑实时性要求,在SRUKF的基础上,引入加性非扩展形式和超球体单形采样,提出了简化超球体平方根UKF算法,通过减少状态维数和采样点的数量,降低算法计算量。建立加性噪声下的单基线GPS/MIMU姿态模型,并采用简化超球体平方根UKF算法进行姿态估计。实验结果表明,单基线GPS的速度信息可以有效提高加速度计对水平倾角的确定精度,姿态估计算法收敛后的最大误差小于0.8°,估计精度与UKF相当,且执行时间仅是UKF的42%。

GPS/MIMU integrated attitude estimation based on simple spherical simplex SRUKF

In view that accelerometers and magnetometer are susceptible to maneuvering acceleration and surrounding environment,an attitude estimation method combining single baseline GPS with MIMU is proposed.The single baseline GPS antenna instead of the magnetometer is used to provide the heading information.The velocity information coming from GPS measurements can be used to compensate the outputs of accelerometers and eliminate the influence of maneuver acceleration on attitude estimation.Based on the square-root unscented Kalman filter,an additive non-augmented unscented transforms and a spherical simplex sampling are introduced to reduce the state dimension and the number of sigma points.Then the proposed algorithm is used to estimate the attitude under the additive noise condition.The results show that the single baseline GPS can effectively improve the horizontal-inclination determination accuracy of the accelerometer.The maximum error by the proposed algorithm is less than 0.8 ° after the convergence,which is equivalent to that of UKF,while the execution time is only 42％ of UKF's.