Research On Damping And Comprehensive Calibration Techniques For Long-Term Inertial Navigation Systems

An inertial navigation systems(INS)is a dead-reckoning navigation system.Such type of navigation system is resistant to disturbances thus can operate properly in any weather and anywhere.However,the navigation errors of an INS also accumulate with time.For long-term inertial navigation systems,which need to work continuously for days,weeks or even longer,measures have to be taken to suppress the accumulation of navigation errors.In this paper,the situation where the external navigation information is not available for most time is considered.In such a research background,damping and comprehensive calibration technologies are studied in detail to improve the performance of the INS.The main contents and innovations of the paper are listed as follows:1.Traditional damping methods suffer from the complex design process and the damping coefficients are hard to change.To overcome the disadvantages of traditional damping methods,a time-variant damping method based on Kalman filter is proposed in this paper.The original critically stable system is stabilized by incorporating output feedbacks.The feedback gains can be calculated through Kalman filter equations,relaxing the engineer from a tiresome try-end-error process.The proposed method endows the system with a big initial damping ratio and a small steady-state damping ratio,which can improve both transient and steady-state performances of the system.Simulation and semi-physical simulation results manifest that the proposed method presents a satisfactory performance.2.The mode switching problem is studied and a mode switching criterion based on the integrated acceleration as well as a compensation method for the mode switching overshoot are proposed.First,the overshoot caused by maneuvering of the vehicle under damping mode is analyzed,which indicates that the overshoot is proportional to the integration of the acceleration.Thus a mode switching criterion based on the integrated acceleration is proposed.Second,the overshoot arising when the INS switches to damping mode is analyzed,which indicates that the overshoot is basically determined by the calculated speed at the switching moment.Thus an overshoot compensation method is proposed.Finally,the overshoot arising under external damping mode is analyzed and a compensation method for mode switching overshoot is proposed.These methods are all verified by simulations.3.A comprehensive calibration method based on constrained Kalman filter is proposed in consideration of the merits and demerits of traditional multi-position and optimal comprehensive calibration methods.Observability analysis of the traditional optimal comprehensive calibration method is first implemented,which indicates the upward gyro constant drift is poorly observable.Then the relationship between the accumulated positioning errors of the INS and the gyro constant drifts is derived based on the working principle of traditional multi-position method.Thereafter,the relationship is integrated into traditional optimal method to constitute a state estimation problem with constraint to the states.Such an estimation problem can be solved by constrained Kalman filter.The resulted comprehensive calibration method based on constrained Kalman filter possesses the merits of both traditional multi-position method and traditional optimal method,i.e.a high precision as well as little time consumption.Comprehensive calibration can arouse system overshoot as well.In this paper,the overshoot is analyzed and a compensation method is proposed.Simulation and semi-physical simulation results manifest that the proposed methods have satisfactory performance.4.The overshoot aroused by comprehensive calibration is studied and the expression of the overshoot is further presented.An overshoot compensation method based on the derived expression is proposed.In consideration of practical implementation,control equations of the INS with the proposed compensation method are presented as well.Simulation and semi-physical simulation results manifest that the proposed method can suppress the overshoot effectively without resulting in extra navigation errors.