论文部分内容阅读
采用高维扩展卡尔曼滤波对多个姿态测量信息进行融合存在求解高维矩阵逆的问题,针对该问题提出了一种基于多源观测信息的低维姿态估计方法。该方法建立了四路适合机动飞行的姿态观测模型,对每路观测信息采用低维混合扩展卡尔曼滤波进行状态估计,并对每路状态估计结果采用马尔柯夫估计进行综合。试验结果表明,该方法可准确估计出角速度常值偏差以保证滤波收敛性,与基于多个测量的高维扩展卡尔曼滤波相比,具有更高的精度。
In order to solve the inverse problem of high dimensional matrices, a high-dimensional extended Kalman filter is used to fuse multiple attitude measurement information. A low-dimensional attitude estimation method based on multi-source observation information is proposed. This method establishes four attitude observation models that are suitable for maneuvering flight, and uses state-of-the-art low-dimensional hybrid extended Kalman filter for each observation information. The state estimation results of each road are integrated by using Markov’s estimation. The experimental results show that this method can accurately estimate the deviation of angular velocity to ensure the convergence of the filter, which is more accurate than the high-dimensional extended Kalman filter based on multiple measurements.