Cooperative Navigation Algorithm for Drone Swarm Combining Robust Adaptive Filtering and Local Feedback in Multileader Mode

In this paper, a cooperative navigation method for multi-leader mode under short-term disturbed or missing condition of measurement information is proposed. The position information correction model of follower drone based on ranging information is derived, and a robust adaptive filter suitable for the short-term disturbed condition of measurement information is designed. In the filter estimation stage, the designed cooperative navigation filter adds a robust factor and an adaptive factor compared with the traditional method. In the filter feedback stage, the proposed localized feedback mode replaces the traditional full feedback mode. This dual approach is employed to suppress the influence of disturbed measurements on the filtered estimation results. In addition, for the short-term missing condition of measurement information, a method of correcting the position information of follower drone by the mean value of the historical estimate of position error state quantity is proposed in the paper. The simulation results show that the proposed method has a greater suppression effect on the positioning error of INSs of different levels when the measurement information is disturbed or failed in the short term. From the results of the flight experiments, the proposed method can ensure the cooperative navigation accuracy of the follower drone in a disturbed environment is better than 6m. Under the condition that the measurement information is unavailable for 10s, the method can reduce the pure inertial error of the INS with the gyro bias stability of 8°/h from 150m to within 35 m.