Improved Adaptive Kalman Filter And Its Application In Transfer Alignment

With the form of change in the modcrn war,more and more airborne precision-guided weapons uses inertial navigation system as navigation equipment,and gradually developed into a main conventional strike force in the modern war.Before launching airborne inertial-guided weapons,how to accurately and quickly achieve the weapon navigation system transfer alignment has become a key technology.For airborne inertial navigation systems transfer alignment problem,several aspects have been studied as follows:Firstly,the relevant terms and definitions of inertial navigation will be introduced,and the basic principle of strapdown inertial navigation system and its calculating method are studied and analyzed;Error equations of SINS are derived through the error propagation mechanism.Then basic theory and characteristics of transfer alignment in SINS are introduced and lever-arm effect between the master inertial navigation systems and missile slave inertial navigation system are analyzed.After that,lever arn effect compensation algorithm is proposed;To the bending deformation and vibration deformation of Wing in transfer alignment are analyzed and the corresponding model is established.Secondly,the application of kalman filtering in transfer alignment has been studied in detail.According to different choose of measurement information in the kalman filter,three classical transfer alignment matching scheme such as the speed matching,attitude matching,velocity and attitude matching were studied in detail.The state space model was deduced respectively.The kalman filtering algorithm has been designed to offer the system of state equation and measurement equation under different matching scheme,and simulated under the two different movement locus.Thirdly,for the problem that the statistical properties of the noise associated with kalman filtering cannot accurately obtain in actual transfer alignment.According to the high dynamic of the SINS working environment and the uncertainty of random disturbance,the Sage-Husa adaptive kalman filtering and strong tracking adaptive kalman filtering were introduced to dynamically update the statistical variance parameters of the system noise.After that,two intelligence optimization algorithm PSO and DE were studied.On the basis of the two algorithm,a new algorithm PSO-DE was introduced based on the two fusion approach.Using the information of kalman filtering to construct fitness function.Finally,Combined the PSO-DE with the adaptive kalman filtering,a new auxiliary adaptive kalman filtering method was presented.the velocity matching scheme for the transfer alignment simulation with the ordinary kalman filtering,Sage-Husa adaptive kalman filter and PSO-DE adaptive kalman filtering were used in the paper.Two different cases are considered respectively:The one is property of system error statistical is constant but not know accurately;the other is statistical feature of system error is change.The simulation results show that under the condition of two kinds of simulation,the estimation precision of the standard kalman filter for some estimates is fall even divergence.Adaptive kalman filter can estimate the related state better,but the estimation precision is volatile.The PSO-DE algorithm and PSO algorithm simulation comparison in transfer alignment.The adaptive kalman filter not only estimate the theoretical value better,but the robustness is better and estimation precision is higher in this paper.