LINEAR ADAPTIVE CONTROLLER FOR CONTINUOUS SYSTEM WITH CONVOLUTION FILTER

摘要

The convolution filter is used for optimal parameter identification and state observation. Two problems are presented. The first one is connected with new improvements in identification method of continuous systems known as "modulating functions". In the primary stage of this method the parameter identification of a continuous SISO model described by an ordinary differential equation is based on convolution of the measured input-output signals and a special filtering function with compact support. This leads to a new model that has an algebraic form with the same parameters. In the next stage we formulate and find the solution of optimal identification problem for the latter model by minimizing the norm of the error in function space L~2. The second problem is synthesis of LQ stabilizing adaptive controller based on the identified parameters and on the state of the system which is concurrently reconstructed by the estimator.