Vibration vision measurement and wavelet neural network control of flexible hinged plate
To address the vibration challenges in flexible thin plate structures like solar panels on space-craft,this study investigates a translational flexible hinged plate system.A binocular vision-based measure-ment and control experimental platform is developed.This platform employs the binocular stereo vision technique for vibration detection,and introduces a self-recurrent wavelet neural network controller(SRWNNC)to mitigate vibration.The system's binocular vision is precisely calibrated.Utilizing the princi-ples of disparity and advanced image processing algorithms,it calculates the three-dimensional coordinates of specific markers to capture vibration signals.A finite element model of the system is constructed,facili-tating the identification of system model parameters.Following this,the SRWNNC is trained within a sim-ulation environment using the identified model parameters,aiming for effective vibration control in the ex-perimental system.Experiments and simulations are conducted on the system,focusing on both fixed base and translational trajectory movements,to evaluate the effectiveness of binocular vision in vibration detec-tion and the SRWNNC in active vibration suppression.The findings confirm that the binocular vision sen-sor achieves a high accuracy less than 0.1 mm in detecting vibrations,and the SRWNNC outperforms tra-ditional large gain PD controllers in damping vibrations,thus validating the efficiency and accuracy of the proposed vibration detection and suppression methods.
国家科技期刊平台
NSTL
万方数据
