The Study Of Evaluation Methods Of AUAV's Hit Precision In Small-Sample Circumstances

Because of special military functions, Attack Unmanned Aerial Vehicle(AUAV) has gained more and more attention from different countries. One of the critical stages of AUAV development is performance estimation, including precision evaluation. For the sake of highly cost of such system, great number of target practices are unacceptable and experiment data are always limited. In such cases, small sample statistic estimation methods are needed in order to perform weapon precision estimation.In this paper, firstly, the features of AUAV are analyzed so as to choose proper precision evaluation method of AUAV. Secondly, the Bayesian method, Bootstrap method and national military standard method are introduced to evaluate weapon precision estimation in detail. Based on the simulation results, the benefits of small sample statistic estimation methods are stressed. Thirdly, a few computation examples are proceeded to examine the effects on Bayesian method and Bootstrap method of different factors. Also the credibility of the prior information is introduced to analyze the uncredibility of prior information. Preliminary relationship between prior information's quantity, credibility and Bayesian estimation result is established with the introduction of different amount of prior information.Finally a Bayesian-based bootstrap estimation method, which combines the advantages of different small sample estimation methods, is proposed according to the specialty of AUAV. With this method, the precision estimation process is finished for the X-type AUAV. Compared with currently used engineering methods, this new method has its own advantages and disadvantages from the point of views of experiment length, the risk of provider and consumer and so on. In the end, the integration assessment of hit precision is designed.