Uncertainty behavior research of hybrid underwater glider

Abstract

Hybrid underwater glider (HUG) combines behaviors of autonomous underwater gliders and those of traditional autonomous underwater vehicles. HUG can dive along a saw-tooth trajectory in the gliding mode and autonomously cruise in the level-flight mode. In the study of HUG dynamics, the mathematical model and the physical parameters of the vehicle are assumed to be deterministic. The simulation results cannot describe the actual motion of the system completely based on the deterministic hypothesis, due to the existence of uncertainty in the real environment. In this paper, parameter uncertainties of HUG are transformed into deterministic problem by some treatment methods. Monte-Carlo and polynomial chaos methods are brought forward to solve the dynamic equations of HUG containing part of probabilistic parameters and the characteristics of these 2 methods are compared. The results show that polynomial chaos method is more effective than Monte-Carlo simulation to achieve the same accuracy. Therefore, polynomial chaos method is chosen to analyze the dynamic uncertainties of HUG. After simulating the different uncertain parameters in the different working models, the results show that the uncertainties of length and diameter of HUG have little effect on the motion performance parameters in the gliding mode and the level-flight mode, but the uncertainties of installation angle of propeller have a great effect on motion performance parameters in the level-flight mode. The present work provides guidance for the actual processing and the improvement of system design.