Aircraft intelligent anti-disturbance control method based on deep reinforcement learning

Abstract

The invention discloses an aircraft intelligent anti-disturbance control method based on deep reinforcement learning, and belongs to the technical field of aircraft navigation, guidance and control. On the basis of a traditional nonlinear anti-disturbance controller, a reference observer network and a deep reinforcement learning method are combined to form an overall intelligent control framework. Then training the reference observer network to enable the reference observer network to establish an aircraft accurate inverse dynamic model, and further generating feedforward control input; and meanwhile, a deep reinforcement learning TD3 algorithm is combined with a traditional nonlinear anti-disturbance controller, and TD3 is used for adjusting control gain parameters of the anti-disturbance controller in real time, so that feedback control input in the overall control framework is formed. Feedforward control and feedback control are combined to obtain an intelligent anti-disturbance control law of the aircraft. The method improves the control performance, adaptability and robustness, and is universal for various types of aircrafts.

Description

technical field [0001] The invention belongs to the technical field of aircraft navigation, guidance and control, and in particular relates to an intelligent anti-disturbance control method for aircraft based on deep reinforcement learning. Background technique [0002] Aircraft (flight vehicle) refers to equipment that flies in the atmosphere or outside the atmosphere (space). It can be divided into several categories such as aircraft, spacecraft, rockets and missiles. It has been widely used in military and civilian fields in recent years. In order to accurately complete increasingly complex, diverse, and precise flight tasks, how to design an attitude control system with excellent control performance has always been an urgent problem to be solved by scholars. [0003] For the problems of strong nonlinearity, strong coupling, parameter uncertainty, parameter time-varying, and external disturbances in the aircraft attitude system, nonlinear anti-disturbance control methods ...

AI Technical Summary

Problems solved by technology

[0004] However, such traditional control methods inevitably have the disadvantages of low adaptability and poor robustness to problems such as complex environments and uncertainties.

Images