关键词: 无人机, 相对导航, 非高斯噪声, 鲁棒自适应滤波, 鲁棒性

Abstract: An adaptive Huber-based cubature Kalman filter (CKF) algorithm with noise estimator is proposed to solve the problem that the measurement noise of vision-based relative navigation sensor for unmanned aerial vehicles (UVAs) formation follows non-Gaussian distribution. The Huber technique based on solving the linear regression problem and the covariance matching method are combined in the proposed algorithm. The residual sequences are used to estimate and tune the statistical characteristics of process noise and measurement noise on line, and then the received measured data are weighted by using the forgetting weighted parameters, thus estimating the relative position, relative velocity and relative attitude information among unmanned aerial vehicles accurately, and improving the adaptive capability of Huber-based CKF algorithm. The simulated results show that the proposed algorithm has strong adaptability to the statistical properties of the contaminated noises, higher estimation accuracy and stronger robustness compared with the standard CKF algorithm. Key

Key words: unmannedaerialvehicle, relativenavigation, non-Gaussiannoise, robustadaptivefiltering, robustness