Model-oriented Experimental Design And Parameter Optimization Of The Missile

Performance analysis and evaluation of the missile equipment is an important part of the equipment manufacture,design finalization,equipment acquisition and operation application,playing an indispensable role in the equipment development,which is also the fundamental guarantee of satisfying the demand of future combat.This paper discusses the missile performance analysis and evaluation based on the theory of experimental design,under the background of model-oriented condition.1)Aiming at the phenomenon that the unscented filtering results may be inaccurate or even divergent in the practice of real-time trajectory estimation,this paper proposes an filter parameter optimization method based on experimental design in order to overcome the this phenomenon.From the perspective of the filter framework,this paper selects the relevant parameters,uses orthogonal design,and conducts the simulation experiments using the measurement data constructed from the theoretical trajectory,then determines the optimal level of influencing factors.The examples show that this method is of high validity,and can improve the robustness and convergence performance of the UKF filter.2)With the illustration of classical regression model and experimental design optimum criterion based on the information matrix,this paper emphasizes on the discussion of a more practical method in experimental design optimization that contains a certain number of supporting points,proposing the maximum information experiment point set design and its algorithms.Furthermore,considering the small sample,inheritance,multi-sources characteristic of the missile tests,this paper puts forward corresponding improvements including:‘re-sift' principles based on ‘cost' and ‘distance',weighted experimental design with the consideration of temporal information weights,additional experimental design based on optimization criterion or extreme values.3)After studying the optimization of experimental design,this paper discusses model-oriented experimental design and parameter estimation problem of the missile.prior information are utilized to describe the partial unknown relationship between the performance index and its relevant factors,therefore create the mathematical response model.This paper further adopts sparse optimization theory to transform the parameter estimation problem of sparsity-included model into a sparsity-reconstruction problem.Based on the ‘design finalization with the sequential experiments step by step' policy,this paper puts forward the sparse analysis-based modeling of combined sequential experimental design to give a model,method and process of the integrated design and evaluation under the consideration of the whole process for the missile flight test,which combines the experimental design of flight scheme and data analysis of the results.4)Methods suggested by this paper are evaluated with specific examples.The results validate that,maximum information experimental point set design tends to enhance the accuracy of the parameter estimation compared with classical maximum entropy design,with a better match with the response surface character.Besides,take the two-step experiments for example,this paper explains the process of the sparse analysis-based modeling of combined sequential experimental design.And it can be concluded that experimental points generated from the above-mentioned model are more consistent with the practical missile experimental regulation,and the expected prediction error of the response function solved by proposed method is smaller.