论文部分内容阅读
使用四元数卡尔曼滤波器进行姿态估计的性能与加速度噪声协方差R有很大的相关性。以四元数卡尔曼滤波器递推计算的新息协方差与测量的新息协方差的差值尽可能小为条件,采用基于新息的自适应算法,在线估计加速度计噪声的协方差R,得到符合实际加速度噪声水平的卡尔曼滤波器增益K,对微小型飞行器进行姿态估计。为了减少由新息引入自适应计算的测量噪声,使用渐消记忆的方法对测量的新息协方差进行处理。仿真结果表明,基于新息的自适应四元数卡尔曼滤波器在加速度计噪声协方差R难以测量时或R变化时,相对于四元数卡尔曼滤波器能更准确地获得姿态的估计,减小了滤波器对模型参数的依赖,增加了滤波器的鲁棒性。
The performance of attitude estimation using quaternion Kalman filter is strongly correlated with the covariance R of acceleration noise. The covariance of the accelerometer noise is estimated online using the adaptive algorithm based on the new information, taking the difference between the new covariance calculated by the recursion of the quaternionic Kalman filter and the measured covariance of the new covariance as small as possible , The gain K of Kalman filter which accords with the actual acceleration noise level is obtained, and the posture of micro-aircraft is estimated. In order to reduce the measurement noise introduced adaptively by the new interest, the measure of interest covariance is processed using the fading memory method. The simulation results show that the adaptive quaternion Kalman filter based on the new information can obtain the attitude estimation more accurately than the quaternion Kalman filter when the accelerometer noise covariance R is difficult to measure or R changes, Reduce the dependence of the filter on the model parameters, and increase the robustness of the filter.