Vibration Attenuation And Attitude Control Of Flexible Aerocrafts Based On PID Wavelet Neural Networks

For the obvious flexible vibration problem in large length-to-diameter rationmissiles, this paper exploringly researches on the vibration characteristics of flexiblemissiles, the effect of structure flexibility on the control system of missiles, and how torestrain the flexible vibration to guarantee attitude tracking. Flexible missiles’ flexiblecharacteristics like large amplitude and long attenuation span, largely interfere thesensors’ measurement in missiles’ control system, which makes the phenomena ofservoelasticity between the missiles’ structure flexibility and control system stand out. Ifregardless of the coupling effect when designing the controller, it will be difficult totrack the missile attitude, even unable to control stably. Therefore, it’s necessary toresearch on the flexible vibration and the vibration restraining methods of flexiblemissiles.The paper adopts simplified Euler uniform beam with tail thrust to model flexiblemissile, deduces the beam vibration equations through bending vibration theory.According to the vibration equations, transverse vibration model of flexible missile isobtained. Combine longtitudinal rigid body model of disturbed short periodic motion,the general linearizational model of flexible missile is built.Based on the above general model, missile’s vibration characteristics are analyzed,and the stability of control system is analyzed as well through theoretic research andsimulation. The result shows that, flexible vibration affects the stability of missilecontrol system, while the increase of thrust makes the control tend to diverge. So, it’sindispensable to design controller to restrain the vibration of flexible missile.Toward the above vibration control demand, PIDWNN(short of PID WaveletNeural Networks) control method is proposed, which is consisted of classic PID controland Wavelet Neural Networks. The study object is controlled by the PID controller in aclosed loop, while WNN can quickly adjust and update PID controller variables on lineaccording to missile motion respond, as WNN has good localized characteristics of timeand frequency, and possesses Neural Networks’ self study ability. Even when largevibrations exist, the controller can weaken or avoid the vibration disturbance in time,which keeps the control performance in best conditions. From the simulation result, it’sapparently that this method not only guarantees the accurate tracking of missile flightattitude, but also restrains the unstable influence brought by flexible vibration, whicheffectively overcomes the coupling between structure vibration and control system.Finally, robustness of the control method is investigated, it shows the bearable uncertainscope of flexible missile model visually, which illustrates the effectiveness of PIDWNNcontrol method in dealing with the vibration problem of flexible missiles when flying inactual complicated uncertain circumstances.