ATR Algorithm Performance Evaluation At Extended Operating Conditions

ATR performance evaluation is becoming one of the aspects of the ATR technology area.The performance evaluation methods of ATR algorithm at extended operating conditions arepresented in this thesis.The significance and background of the research have been introduced in the introduction,so as the history and the situation of the evaluation field. The introduction also gives out theoutline of the thesis.ChapterⅡsummarizes ATR algorithms from two aspects of radar target recognition. Therelationship of ATR system performance and ATR system cost and operating conditions aregiven. At last, three evaluation methods are introduced by analyzing the three performancemeasures of ATR.ChapterⅢdesigns the ATR algorithm performance evaluation experiment at extendedoperating conditions and discusses each process of the experiment. The target models are settedup by 3D MAX. The three factors of extended operating conditions are targets, environment andsensors. The content of each factor is analyzed. Some factors are selected and arranged in pairsfor the evaluation experiments at extended operating conditions.The echo of each target atextended operating conditions is simulated. Finally logistic regression model are analyzed and anevaluation method based on logistic regression model is raised.ChapterⅣsets up the logistic regression models of the targets between factors and theperformance of the ATR algorithm.and shows the goodness of fit between the models and data.The last chapter summarizes all the research work of the thesis, draws the conclusions andgives some recommendations for further research.