非方广义随机系统的状态估值器控制理论与控制工程专业论文.docxVIP

4岭‘4··11 4 岭 ‘ 、 .4』，，， 】I 】I l l l J1 Y1III 19r 4r 0fl 12II 18I 19II 黑龙江大学硕士学位论文 Abstract Witll the continuing development of engineering technology,many achievements about generalized systems have emerged．And those consequences which have been widely recognized by academic and engineering field have been used in different aspect of the modern science．Nevertheless，the engineering technology is becoming more and more complicated,it is difficult to describe the control systems by normal ◆ systems or simple generalized systems．To solve this problem more easily,researchers are begin,ing to focus on the non-square generalized systems gradllally．Non-square generalized systems are the extension of the generalized systems．Under certain conditions，many of the researching achievements of generalized systems Can be transplanted into the non-square generalized systems． Kalman estimator has the advantages of optimality and simplicity,tractability and robustness，and for that reason it has been widely used in the state tracking and estimation areas．The emphases and direction of this dissertation is the combination of the non-square generalized systems and the classical Kahilan filtering theory． Based Oil classical Kalman filtering theory,this dissertation mainly discusses the problem of state estimation of Kalman for non-square generalized systems．The main research results of this paper can be given as follows： Firstly,under the conditions of the regularizability and the complete observability, non-square generalized systems call be transformed into normal systems．Using the classical Kalman filtering theory,we have obtained the Kalman filter,predictor and smoother for non-square generalized systems． Secondly,under the conditions of the regularizability and the strongly conrollabilllity of non-square generalized systems，the non-square generalized systems are transformed into normal systems by using singular value decomposition of matrix ．II· a