Robust adaptive optimal control-based pilot manipulation behavior description method

Abstract

The invention discloses a robust adaptive optimal control-based pilot manipulation behavior description method. The method includes the following steps that: an ROCM model is constructed to augment a controlled object; the index function and model parameters of the ROCM model are determined; pilot optimal control gain is calculated; equation discretization is controlled; robust adaptive state estimation is carried out; unmodeled dynamic estimation is carried out; the pilot optimal control gain is corrected; and pilot manipulation vectors are solved. According to the robust adaptive optimal control-based pilot control behavior description method of the invention, characteristics of memory finiteness and high adaptation of human beings are combined, an improved manipulation noise and observation noise model is adopted, robust adaptive filtering theories are introduced, and therefore, the shortcomings of a traditional pilot model can be overcome, and characteristics of human beings such as memory finiteness, unknown environment adaptability, manipulating behavior robustness and coupled manipulation can be embodied.

Description

technical field [0001] The invention relates to a method for describing a pilot's manipulation behavior when the dynamic characteristics of an aircraft change under an unknown environment disturbance. Background technique [0002] At present, aircraft pilot models in engineering field include McRuer model, optimal control model, pilot structure model, neural network model and so on. Among them, the McRuer model and the pilot structure model are single-channel models, and decoupling processing is required to describe the coupling manipulation; the neural network model faces the contradiction between knowledge experience and online learning, and when the dynamic characteristics of the aircraft change, the original mapping structure may be destroyed; Although the optimal control model (hereinafter referred to as OCM) can perform coupled manipulation analysis, it uses the Kalman filter for state estimation, which does not reflect the gradual adaptation process of humans in an un...

AI Technical Summary

Problems solved by technology

Among them, the McRuer model and the pilot structure model are single-channel models, and decoupling processing is required to describe the coupling manipulation; the neural network model faces the contradiction between knowledge experience and online learning, and when the dynamic characteristics of the aircraft change, the original mapping structure may be destroyed; Although the optimal control model (hereinafter referred to as OCM) can perform coupled manipulation analysis, it uses the Kalman filter for state estimation, which does not reflect the gradual adaptation process of humans in an unknown environment; with a given signal-to-noise ratio, iteratively solves the noise variance, It is equivalent to the pilot evaluating the noise characteristics based on the future that has not yet occurred, which has some contradictions with the pilot's actual thinking process and driving behavior. In addition, this model cannot be applied when the dynamic characteristics of the aircraft change, and the robustness is poor

Therefore, the pilot model in the existing engineering field is not suitable for analyzing the coupled control behavior of the pilot when the dynamic characteristics of the aircraft change under unknown environmental disturbances

Images