Multi-Unmanned-Aerial-Vehicle Navigation Algorithm Based on Dynamic Adjustment

In multi-UAV cooperative navigation, the traditional Extended Kalman Filter (EKF) algorithm uses a fixed noise covariance matrix, which makes it difficult to adapt to dynamic environmental changes and affects the accuracy of state estimation. To address this issue, this paper proposes a Covariance-adjusted EKF (C-EKF) algorithm based on dynamic noise covariance adjustment, which utilizes observation residuals to dynamically update the noise covariance matrix. For process noise, weighted averaging and historical weighted averaging methods introduce enhancements to system adaptability. For observation noise, weighted smoothing and residual correction methods are employed to improve the accuracy of measurement updates. Simulation results show that compared to the original algorithm, the proposed C-EKF algorithm reduces the longitude error from 5.02 m to 2.61 m and the latitude error from 5.02 m to 2.77 m.