在无人机飞行参数测量中,受环境、干扰或自身稳定性等因素的影响,往往使测量数据中存在野值。为提高飞行数据处理精度,提出了一种基于观测器/卡尔曼滤波辨识(OKID)的新算法用于估计飞行参数。该算法直接利用输入输出数据与系统矩阵得到观测器增益,而观测器增益又收敛于稳定状态下的卡尔曼滤波增益,利用卡尔曼滤波增益即可估计飞行参数。Matlab仿真表明该方法能有效地消除飞行数据中的野值。 In the measurement of UAV flight parameters, there are often wild values ​​in the measurement data due to the influence of environment, disturbance or self-stability. In order to improve the accuracy of flight data processing, a new algorithm based on observer / Kalman filter identification (OKID) is proposed to estimate flight parameters. The algorithm directly uses the input and output data and the system matrix to obtain the observer gain, and the observer gain converges to the Kalman filter gain under steady state. The Kalman filter gain can be used to estimate the flight parameters. Matlab simulation shows that this method can effectively eliminate the outliers in the flight data.