Research On A Ring-type Multi-DOF Ultrasonic Motor And A Visual Tracking System Driven By IT

Usually,the target tracking system adopts the electromagnetic motor as the actuator.And the structure of it is a multi-axis PTZ structure.In between technology,smaller,more accurate and faster target tracking system is needed.However,due to the structure of the traditional target tracking s ystem,it is difficult to miniaturize it and since it uses the electromagnetic motor as the actuator,it is a big challenge to realize a lower speed and a higher accuracy.Ultrasonic motor as a newly developed actuator exhibits a lot of advantages,such as small volume,high accuracy and fast responsiveness.Due to the advantages of the ultrasonic motor,it has been widely used in industrial areas.This paper will develop a new multi-DOF ultrasonic motor that can be used in the target track system.First,the motor’s output characteristics and the closed-loop control characteristics have been investigated.Then,according to the previous investigation,the motion control modes of the target tracking system driven by the designed motor are investigated.Finally,based on the previous experiments,the tracking accuracy of the designed tracking system is measured.Firstly,by analyzing the research status of the multi-DOF ultrasonic motor,a new type of multi-DOF ultrasonic motor with four feet driving simultaneously was found.Based on the new multi-DOF ultrasonic motor,an improved motor was designed.The newly designed motor adopts only two slices of ring-typed piezoelectric ceramic to excite three vibration modes which makes the motor smaller,simpler and easier to manufacture and assemble.Then,the finite software ANSYS was used to complete the analysis of the designed motor,including the mode analysis used to degenerate the mode frequency between different vibration modes,the harmonic analysis us ed to compute the electromechanical coupling efficiency and the mechanical quality factor,the transient analysis used to verify the working principle of the motor.Then,in order to finish the control of the target tracking,first,the theory of the particle filter tracking algorithm was introduced.Then,the target tracking and the extraction of the target position program was designed.In order to decrease the computing time of the program while tracking multiple targets,a parallel computing procedure was designed.A self-adaption algorithm was alsoadopted in the program to overcome the change of the environment light illumination.Besides,this paper developed an angle measure algorithm that can measure the rotation angle of the motor in three axes by using a camera as the sensor.This measure algorithm was used in the research of the target tracking control.Finally,the comparison experiments of the motor control and the target tracking were implemented.First,the motor’s closed-loop control characteristics under different control variables were studied.And the results show that the control linearity of the voltage control variable is better than the frequency control variable.But while using the voltage as the control variable,there is a speed dead-zone when the control speed is in low values.However,the frequency control variable exhibits a short rise-time while tracking a step speed signal with a low speed value.Then,the control characteristics under different control algorithms were also investigated.The investigation results show that the PID controller with dead-zone compensation can solve the speed death-zone problem completely.But while using the PID controller with dead-zone compensation to track a sine speed wave,the tracking error is a little bit big.Meanwhile,the variable structural PID controller with dead-zone compensation can not only solve the problem of dead-zone,but also receives a small tracking error in tracking the sine speed wave.The target tracking characteristics under different motion control modes were investigated.The results show that compared with motion synthesis control mode,the motion interpolation control mode receives a smoother tracking curve and a smaller coupling angle.Finally,the target tracking experiments were implemented and the results show that the tracking system obtains a tracking error of 0.5° while tracking a sine angle wave with the frequency of which is 1Hz and the amplitude of which is 10°,a rise-time of 0.22 s and a tracking error of 5 pixels while tracking a step moving target and a tracking error of 34 pixels while tracking a moving target with a speed of which is 32pixel/s.