Robust constrained kalman filter algorithm considering time registration for GNSS/Acoustic joint positioning: Fid-Bau Portal

Robust constrained kalman filter algorithm considering time registration for GNSS/Acoustic joint positioning

Kuang, Yingcai / Lu, Zhiping / Li, Linyang / Chen, Zhengsheng / Cui, Yang / Wang, Fangchao

Abstract Currently, the precise positions of seafloor control points are very important for expanding marine navigation activities. Equipped with a variety of sensors, a survey vessel measures and integrates heterogeneous data to obtain the coordinates of seafloor control points. However, the existence of time error factors such as inconsistent sampling intervals or signal transmission delays will cause time mismatches in multisource time series data. This will greatly reduce the positioning accuracy of seafloor control points. Previous studies have generally ignored the problem of time registration. To address this issue, this paper first briefly introduces the basic principle of positioning a single seafloor control point by using multiple types of sensors. The influences of various time errors on the resolution are analyzed through formula derivations and simulation experiments. Then, a robust constrained Kalman filter (RCKF) algorithm considering time registration is proposed. Based on the Kalman filter algorithm, the RCKF method takes the transmission delay error as a parameter and introduces priori constraints and the robust estimation. So the time registration and the determination of the seafloor control point coordinates are carried out in real time. The performance of the RCKF method in the designed simulation experiment and a testing experiment near Jiaozhou Bayou is discussed. The positioning accuracy is 3.07 cm in simulation when adopting the RCKF method, which is better than other comparison methods. The positioning accuracy of the testing results can be greatly improved from 1.348 m in the common method to 0.12 m based on the RCKF method. The results show that the newly proposed method ensures the consistency of multisource time series and has good stability and accuracy in positioning a single seafloor control point.