To address the problem of significant degradation of global positioning system(GPS)positioning accuracy for unmanned aerial vehicles(UAVs)in complex weather and non-stationary noise environments

an adaptive extended Kalman filter(EKF)algorithm integrating cross-layer assistance of received signal strength indication(RSSI)and correction of normalized innovation squared(NIS)is proposed.The algorithm realizes a priori prediction of environmental interference through cross-layer perception of communication-layer RSSI features

and accomplishes dynamic identification of observation

outliers combined with NIS statistical detection.The dual mechanisms collaboratively realize real-time adaptive adjustment of the measurement noise covariance matrix

R

.The experimental results show that on the STM32F103 hardware platform with a main frequency of 72MHz

the single update time of the algorithm is only 27.1μs

which can meet the stringent real-time requirements of the UAV flight control system.Compared with the standard EKF algorithm

the proposed algorithm achieves a 59.41% improvement in average positioning accuracy in dynamic trajectory tracking and simulated adverse weather interference experiments

and its positioning error characteristics approach the Cramer-Rao lower bound(CRLB).This method incurs no additional hardware cost and can significantly improve the positioning robustness and environmental adaptability of UAVs in signal-degraded environments.