PID parameter design method based on path integration

Abstract

The invention provides a PID (Proportion Integration Differentiation) parameter design method based on path integration. The method comprises the following steps: designing an updating formula of PID controller parameters; the method comprises the following steps: by taking a current PID control parameter as an initial value, generating N exploratory control parameter combinations in a mode of adding limited random noise to the PID control parameter in a current state of an aircraft; respectively controlling the attitude of the aircraft by using the N control parameter combinations, counting each optimization index, calculating a return value, and updating PID control parameters according to a designed updating formula; and taking the PID control parameter obtained at the previous optimization moment as an initial value, optimizing the PID parameter again under the state of the aircraft at the current moment, repeatedly executing, and obtaining the optimal control parameter of the full-trajectory PID controller at the ground design stage. The invention provides a PID controller parameter design method based on data driving, which gets rid of dependence on experience of designers.

Description

technical field [0001] The invention relates to the technical field of aircraft intelligent control, in particular to a PID parameter design method. Background technique [0002] Traditional reinforcement learning algorithms use Bellman equations and neural network approximation techniques to learn optimal strategies, but the learning efficiency is low and the learning effect is not good in systems that need to perform continuous tasks such as aircraft. [0003] In order to improve learning efficiency and obtain the optimal solution, a reinforcement learning method based on path integrals can be used. However, when designing the flight controller of the aircraft, the PID control parameters are often determined at the ground design stage. During the design, the typical feature points are first determined according to the flight trajectory, and the aircraft model parameters are determined at the selected feature points. The designer then determines the PID according to the mod...

AI Technical Summary

Problems solved by technology

However, when designing the flight controller of the aircraft, the PID control parameters are often determined at the ground design stage. During the design, the typical feature points are first determined according to the flight trajectory, and the aircraft model parameters are determined at the selected feature points. The designer then determines the PID according to the model. The control parameters of the controller rely heavily on the experience of the designer in the design, and the PID parameters designed by different designers are quite different, which affects the control effect of the controller

Images