Compensation of time-varying clock-offset in a LBL navigation

Abstract

This paper presents compensation of the clock-offset in a long baseline (LBL)navigation. It departs from the existing literature mainly in dealing with a time-varyingclock-offset, i.e. the clock-rate drifts over the time. Specifically, the clock-offsetdynamics are introduced to the ToFs as an autoregressive filter.Subsequently,interactions among the now biased ToFs and the kinematics of an autonomousunderwater vehicle (AUV)–the navigation subject–are represented in a state-spaceform. Implementing the so-called graphic approach, minimum sensor requirementfor this system’s observability is then explicated. Finally, a standard discrete Kalmanfilter is deployed as the state estimator. By simulation, it is demonstrated that theestimator manages to compensate the offset and to provide localization with less than1 m accuracy