Research And Application On Two Degree Of Freedom Pan-Tilt Visual Servo System

Moving target tracking based on the visual information feedback is a research hotspot in the field of computer vision.The basic principle is to detect and identify the moving target from the video sequence acquired by the camera,extract the image feature information of the target and feed it back to the Pan-Tilt.By designing the controller,the PTZ is controlled to rotate in the horizontal and pitch directions to capture the target such that the target always appears in the center of the camera plane.At present,most of the visual servo control system uses the classical control theory algorithms,such as PID controller algorithm.In the case of control accuracy is not high,by adjusting the PID parameters can roughly meet the system requirements.However,the two-degree-of-freedom PTZ visual servo tracking system in this paper has high requirements for system control accuracy,and the traditional PID controller has its own technical bottleneck and can not take into account the rapidity and stability of the system at the same time.In recent years,control method based on modern control theory has been much riper,the design of state feedback based on the state space model can optimize the performance,cost and time of the system under given constraints.First,this paper starts from the actual experiment,build a two-degree-of-freedom PTZ visual servo system,and load it on the mobile robot platform;Second,the pseudo-random signal sequence is used as the identification signal.The Hankel matrix of the system is used to determine the order of the system.Then Subspace based State Space System identification method and the Least Squares identification method are used to systematically identify the model of the PTZ visual servo system,The model is validated according to the response of the actual system.The results show that the identification model coincides with the actual system;Then,to meet the requirements of controller design in practical engineering,the order of the system model needs to be reduced;Third,the PID controller is designed based on the classical control theory.Then,based on the modern control theory the LQR output feedback based on the internal model principle is designed to improve the fastness,stability and tracking precision of the system.The bilinear transformation method is used to discretize the designed controller;consider the camera ring with the closed inner ring as a new system,then the similar identification method can be used to get the transfer function of the new system and controller can be designed to control the whole system.Finally,the control algorithm is designed on the experimental platform.The experimental results show that the proposed control scheme can ensure that the system has higher tracking precision and faster response speed.The comprehensive performance is far superior to the traditional PID controller.