| For automobile robot, self-localization is a fundamental ability. Inertia Navigation System is a self-localization way and is not able to substitute in the positioning system of the mobile robot in unknown environment. However, the error caused by the drift of the inertia components is the neck of bottle to accomplish positioning. It is the main task for this thesis to analyze and to model the FOG's drift, the influence of the drift to positioning and how to restraint the drift.The positioning system based on multi-sensor information fusion has been introduced in this thesis. The principle of dead reckoning system has been described. The positioning error caused by the drift of the inertia components has been analyzed. The Kalman filter is an optimal way to restraint the error of Inertia Navigation System, building the error model is the preliminary condition to realize.After introducing the FOG's principle and the cause of the FOG's drift, the time series is adopted to analyze the drift data of E.core1000RD FOG and it is concluded that (1) the FOG's drift is a non-stationary time series and includes random component and constant component, but it is a stationary time series as the temperature is stable; (2) the error model of the FOG's drift is time-varying and the constant component of the FOG's drift has a relation of nonlinear to the temperature; (3) the error model of the FOG's drift is expressed as AR(1); (4) at last, combining with the character of the robot working, RLS is adopted to adjust the model parameter online, the one-step-predicted-output is used to confirm the availability of the model.It is a necessary condition to construct the FOG's error model for estimating and diagnosing the fault of sensors and achieving positioning system based on multi-sensor information fusion. |
PDF Full Text Request