Greedy orthogonal least square method and parameter and time lag identification method based on same

Abstract

The invention discloses a greedy orthogonal least square method and a parameter and time lag identification method based on the greedy orthogonal least square method. A parameter vector in an identification model is converted into a high-dimensional sparse parameter vector by using a parameterization method, and a greedy criterion and a calculation mode are simplified by using Householder transformation to avoid matrix inversion; and the sparse identification method is used for identifying the position and the numerical value of the non-zero element in the sparse vector. The method can be usedfor carrying out joint estimation on the time lag and parameters of the system under limited sampling data, the identification cost is reduced, and the precision is high.

Description

technical field [0001] The invention relates to the field of system identification, in particular to a greedy orthogonal least square method and a parameter and time-delay identification method based on the method. Background technique [0002] Many practical systems in the industrial production process (such as thermal systems, chemical processes, sensor networks, wireless communication systems, etc.) contain unknown time delays, and the existence of unknown time delays will make the output of the system unable to respond in time The stability becomes poor, thereby affecting the control of the system. If the system model is known, the parameter identification of the system is the premise of controlling the system. System identification, control theory, and state estimation are three closely related research fields in modern control theory. Control theory and state estimation cannot be separated from the mathematics of the system. Therefore, identification is the basis of c...

AI Technical Summary

Problems solved by technology

CS theory is an important theory for the study of sparse systems. It can reconstruct unknown high-dimensional sparse signals based on low-dimensional observation data and measurement matrices. In many fields, sparse systems have been studied in depth based on this theory. However, in the control There is not much theoretical research

The CS theory shows that: under certain conditions, the sparse system can realize the estimation of the system parameters when the amount of sampled data is lower than the dimension of the system parameters. Commonly used compressive sensing reconstruction methods include convex optimization methods and greedy algorithms. Its superior performance and easy implementation have made great progress in recent years. The Orthogonal Matching Pursuit (OMP) algorithm is a typical greedy algorithm. It can obtain accurate reconstruction of the signal when there is no noise interference. On the basis of the OMP algorithm, a compression algorithm is also proposed. A series of improved algorithms such as sampling matching pursuit (CoSaMP) and regularized orthogonal matching pursuit (ROMP). The problem with this type of algorithm is that the algorithm needs to calculate the inverse matrix to find the least squares solution

Images