Research On LQ Control Method Of Photoelectric Platform

Gyro stabilized platform is an important part of the automatic control system ofairborne targeting pod，it is the main component of line-of-sight (LOS) stabilization,the platform can isolate the vibration of the loading machine, and follow the drivingcontrol instruction to complete the stable tracking of targets, is a complex systemwhich integrates light, machinery, electricity in one. On the basis of theory andexperiment, effective control strategy of a photoelectric platform is researched. Thispaper adopts linear quadratic Gauss control (LQG control) to restrain frictioninterference and interference suppression of the detector.This paper ifrst introduces the history and development process of AirborneOpto-Electronic Platform, then introduces the traditional control method of thephotoelectric platform. And then introduces the basic principles and superiority ofoptimal control and linear quadratic optimal control.This paper analyzes the structure and the principle of isolation of airborneoptoelectronic platform, established the model of rate gyro, the model of motor andframework, the model of interference source, the model of control loop, and analyzesthe torque rigidity. This paper analyzed all the interference factors which inlfuencesthe system stability precision, the friction force is the main interference factor, inorder to be able to accurately compensate the friction disturbance, the frictionmodel should be identifying ifrstly. This paper compares several different frictionmodels, and ultimately selected one for the actual platform model. This paper analyzes the principle and characteristics of the linear quadraticGauss control (LQG control), an important characteristic of the linear quadraticcontrol is making the optimality and stability together, it can realize the optimalityand stability of the system at the same time. The linear quadratic Gauss controlsystem (LQG) a neutral system which meeting the conditions of the separationtheorem, the parameter identiifcation, state estimation and controller design can beindependent of each other. To the system, in which the model uncertainty and theenvironmental uncertainty is existing at the same time, LQG can achieve the optimalsolution, an have a strong inhibitory effect on noise.The core problem of linear quadratic optimal control is how to the two weightmatrix Qc and Rc, the traditional methods mainly include trial and pick bysimulation, the numerical iteration method and the analytical method. In recent years,many intelligence algorithms are applied to select the two control weight matrix ofthe linear quadratic, such as genetic algorithm. Based on the traditional geneticalgorithm, this paper will design a kind of improved genetic algorithm and use it tooptimize the weight matrix.At the end of this paper, the experimental analysis of the above theory andalgorithm is showing, the results have made consistent with the research and designof the theoretical, the correctness and engineering practical value of the theorydesign is proved.