Development of synthesis method of fault tolerant systems for autonomous underwater robots with navigation sensors failures

The development of a synthesis method of highquality fault tolerant systems for autonomous underwater robots with navigation sensors failures is addressed in this paper. The offered new method bases on using of kinematic models of these robots and a special fusion of data obtained from their onboard navigation sensors. The advantage of method is simplicity of realization and high precision of compensation of detected faults in the conditions of uncertainty and essential variability of parameters of environment and underwater robot. The results of modeling confirm high efficiency of functioning of the synthesized fault tolerant systems.

I. Introduction

Today the autonomous underwater robots (AUR) are the most effective research instrument of the World Ocean, and performance of various underwater and technical works. The navigation sensors are an important element of underwater robots on which their ability to carry out objectives depends. Information from these sensors are used for formation of trajectories of the motion of robots. Refusal or failure in operation of any navigation sensor leads to errors in performance of objectives or even to loss of AUR. Therefore, the problem of timely detection of faults in these sensors is topical. The important task leading to increase of efficiency of AUR at emergence of insignificant faults at which else it is possible to execute and successfully to finish operations (mission) is formation of special control signals, which allow to automatically saving the most important characteristics of robots during their work. The solution of this task is provided by fault tolerant systems.