Studies On Sins Calibration Performace Quantitative Evaluation Approach Based On Bayesian Smoothing

Inertial navigation technology gains significant development in recent years,which comes as advanced initial alignment approach and comprehensive calibration method.It is obviously that exhibits efficient improvement on the performance of strapdown inertial navigation system(INS).Consequently,INS performance evaluation approach demands to assess its key techniques,including alignment and calibration.Moreover,the evaluation trend shifts from indirect evaluation with final navigation error to directly quantified evaluation with relevant key parameters in alignment or calibration.We focus on the comprehensive calibration and alignment technologies aim to develop their accuracy assessing approaches,to evaluate the misalignment angles of initial alignment and the constant drifts of calibration,respectively.Bayesian estimation theory,which is the primary algorithm in INS performance evaluation,are studied to analyze the calculation property and performance of Bayesian smoothers.Analysis result proves that Bayesian smoothing algorithm is appropriate for alignment and calibration performance evaluatioin.Furthermore,we modified the calculation framework of fixed interval smoother(FIS)and simplified linear/nonlinear model based nonlinear FIS to improve the computational efficiency of FIS.Initial alignment performance evaluation based on Bayesian smoothing is researched from following aspects.We applied fixed point smoother(FPS)and FIS into the evaluation problem,and precisely assess all misalignment angles.Observation analysis is introduced to indicate the relation between the validity of alignment performance evaluation scheme and the observability of misalignment angles.Then,we design two evaluation approaches which employ accurate azimuth reference and large acceleration maneuvering to precisely assess the azimuth misalignment angle.Meanwhile,we present an alignment accuracy evaluation approach for attitude matching transfer alignment,in which master INS and slaver INS are mounted flush to the same slab and establish azimuth installation error reference.Large azimuth misalignment performance evaluation method is researched by applied nonlinear smoothers.Furthermore,in pseudo inertial coordinate,the corresponding evaluation model is established as well as the transform and compensation method of reference informance are derivated to assess the alignment accuracy in pseudo inertial coordinate.Finally,we focus on the performance evaluation of comprehensive calibration,in which the key problem is to estimate gyroscope constant drift.Theoritic analysis indicts comprehensive calibration takes hours fitering.Therefore,system level damping technology is introduced to damp navigation oscillation errors.A novel level damping approach is designed by merging the damping differential equations into navigationg error equations.Then,the proposed filter based damping model is established and applied in the calibration of sigle-axis rotation modulation INS.The azimuth gyroscope drift is precisely estimated and reset together with the other navigating errors.For the entire calibration performance evaluation,accurate attitude information assisted gyroscope constant drift estimation approach is presented.The INS attitude errors are compensated,then establish combined observations to conduct the assessment of gyroscope constant drift by smoothers.