Hybrid state estimation method for high-dimensional coupling uncertainty system

Abstract

The invention provides a hybrid state estimation method for a high-dimensional deep coupling uncertain system, which comprises the following steps of: firstly, constructing a state, parameter and measurement model of the high-dimensional deep coupling uncertain system, and designing a corresponding observer model to obtain estimated values of the state and the parameter; secondly, discretizing thesystem to decompose the system into a low-dimensional discretized hybrid model, and further obtaining the low-dimensional discretized hybrid model and a parameter model; and finally, taking the observer output estimation value as an auxiliary signal, performing filtering processing on the state estimation value of the low-dimensional discretization hybrid model by using a volume Kalman filteringalgorithm, and outputting a state value of the low-dimensional discretization hybrid model. According to the method, the system model is corrected through the estimated value output by the observer, the system state estimation precision can be effectively improved on the premise that the system stability is guaranteed, meanwhile, the calculation dimension of the filtering calculation process is reduced through the low-dimensional volume Kalman filtering algorithm, and the method is suitable for high-dimensional, coupled and nonlinear system state estimation with uncertainty.

Description

technical field [0001] The invention relates to the technical field of mixed state estimation, in particular to a method for estimating a mixed state of a high-dimensional coupled uncertain system. Background technique [0002] For linear exact models, the linear Kalman filter algorithm provides the optimal recursive solution of the system. But for the actual complex engineering system, the system model is usually nonlinear, high-dimensional, coupled, and the system model has parameter uncertainty. At this time, the classic model-based estimation method will have problems such as excessive calculation, reduced estimation accuracy, and even divergence. At present, there is no effective state estimation method for high-dimensional, nonlinear, and uncertain systems. Therefore, the development of state estimation methods under high-dimensional, nonlinear, and uncertain system models has certain cutting-edge and practical application value. Contents of the invention [0003]...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the technical problems of large amount of calculation, low estimation accuracy and divergent results in the existing state estimation methods, the present invention proposes a hybrid state estimation method for high-dimensional coupled uncertain systems, which can realize high-dimensional, nonlinear, uncertain nonlinear Accurate, effective, and stable state and parameter estimation of linear systems, reducing the computational dimension of the estimation process, and improving the accuracy of system state estimation on the premise of ensuring the stability of the estimation process

Images