An adaptive window compensation method for ranging outliers in the slave node assisted by the master node

Abstract

The slave node navigation, supported by the master node, involves the slave node (equipped with low-precision navigation instruments) periodically receiving positioning data from the master node (equipped with high-precision navigation instruments) through the acoustic sensors. The acoustic sensors also measure the relative distance between the two nodes and uses the Extended Kalman Filter to correct the slave node’s positioning error, thereby improving its positioning accuracy. However, the complex underwater acoustic environment may cause packet loss or ranging outliers in inter-node communication, degrading the slave node’s navigation performance under master node assistance. In order to solve the above problems, this paper proposes an adaptive window compensation method for ranging outlier in the slave node assisted by the master node. Firstly, in underwater acoustic communication between master and slave nodes, the current ranging outliers are detected by utilizing the secondary residual derived from the one-step prediction residual and the state estimate at the previous information interaction time. Secondly, the length of the adaptive window for compensating anomalies is determined based on the relative azimuth angle between the master and slave nodes, as well as the change in the azimuth angle of the slave node between adjacent information interaction times. Finally, the window weights are calculated using the slave node’s travel distance during the adjacent information interaction periods, thereby completing the compensation for the relative distance between the master and slave nodes at the current information interaction time. To validate the proposed method, simulation results show that the average positioning error of the slave node after ranging anomaly compensation is 6.73 m, which is comparable to the error (6.76 m) in scenarios without outliers. Further validation through field tests on a lake demonstrates the method’s effectiveness: the average positioning error is reduced from 17.05 m to 3.53 m after outlier compensation. Both simulation and experimental results confirm the efficacy of the proposed approach.