Comparative analysis of recursive estimation algorithms for AUV collaborative navigation problem in the case of abnormal outliers in measurements

When solving the autonomous uninhabited underwater vehicles collaborative navigation problem, the following algorithms were compared in terms of accuracy, consistency and computational complexity: the classically constructed extended Kalman filter, the extended Kalman filter, additionally using the procedure for rejecting abnormal emissions in measurement errors and the maximum correntropy Kalman filter. The comparison was carried out under conditions of measurement noise, both Gaussian and non-Gaussian, containing pulse-type interference.