The Study On Visual Ball&Plate System Motion Control Based On FOA

Ball&plate system is a under-actuated,strong coupling,multivariable and nonlinear kinetic system,it extracts the ball position from image or touch screen,then gives feedback to the controller,controller controls the rotation angle of the motor to adjust the tilt angle of the flat,thus realized positioning control,trajectory tracking control,winding control of the ball.Research of ball&plate system involves mechanical structure designing of the experiment platform,system kinetic analyzing and nonlinear modeling,modern control theory,intelligent control theory,digital image processing,trajectory planning and tracking,and many other fields.Ball&plate system becomes a typical control object in control theory research,due to the complexity of itself,universality of involved fields,and simplification of physical model,it can test availability of various control algorithm,the research results of theory analysis and control method based on it can promotion to robot control,satellite positioning,ship navigation and other practical engineering issues.Therefore,this paper selects visual ball&plate system as the research object,analyzes and studies the existing control method FUZZY-PID controller of the system,optimizes the visual inspection scheme,and aims at the problem that the adjusting time is longer and the steady-state error is larger in FUZZY-PID,this paper introduces the newer and more efficient algorithm,Fruit Fly Optimization Algorithm(FOA)to tuning the PID parameters,proposes a modified Variable-Step FOA(VSFOA)to realize dynamic tuning PID control parameters on-line,then designs VSFOA-PID controller,realizes positioning control and trajectory tracking control of the ball,experimental results show,when using VSFOA-PID controller,positioning control accuracy is higher,adjusting time is shorter,and trajectory tracking is faster.The main work of this paper is as follows:Firstly,this paper expounds the research significance,research background and situation of the ball&plate system and FOA,introduces the physical structure and working principle of the ball&plate system,analyzes and studies the theory,and builds the simulation model of the system based on using the Lagrange principle to establish the mathematical model of the ball&plate system,selects flat tilt angle as control volume to simplify the model,gets the decoupling linear model on x/ y axial at last.Secondly,because of the noise generated in image transmission process,changes of the light,the ball cannot be positioned accurately,so this paper denoises the image at first,then using OTSU to obtain the binarization threshold of the grayscale image,thus to accurately segment the ball image,and extract the ball centroid.Since there have a wide background in the image,it will take a lot of calculation time in image processing,so this paper proposes to divide the image to background area and target area,the image of the ball and plate are consist in target area.The optimized visual inspection scheme proposed in this paper,improves the positioning accuracy,and enhance the robustness of system visual inspection.Thirdly,due to the process is complex and the adaptability is poor,traditional PID parameters tuning on-line is difficult to meet the requirement of fast,accuracy and stable,this paper introduces FOA to tuning the PID parameters on-line,thus improve the control quality of the system.The step size in general FOA is a fixed value,it makes FOA cannot balance the global search ability and local search ability,search accuracy and convergence speed,so this paper improves the general FOA,proposes VSFOA in which the step size reducing decrease with the increasing of iteration times,so the algorithm has high global search ability and fast convergence speed at the initial stage of iteration,and high local search ability and search accuracy at the later stage of iteration,then the VSFOA-PID controller is built based on using VSFOA to dynamic tuning PID parameters on-line.This paper using PID controller,FOA-PID controller,VSFOA-PID controller to do positioning control simulation experiment on ball&plate system model,the experimental results not only verify the effectiveness of the VSFOA,but also indicate that VSFOA-PID controller can shorten the adjusting time,reduce the overshoot and the system error performance index of PID controller sharply,when compare with FOA-PID controller,VSFOA-PID controller has shorter adjusting time and lower system error performance index,so it can be used to realize more complicated control,such as trajectory tracking control and winding control;then using PID controller,VSFOA-PID controller to do trajectory tracking control,the experimental results indicate that VSFOA-PID controller has shorter system response delay,smaller error performance index.Finally,using the VSFOA-PID control algorithm in real-time control system,achieve the positioning control and trajectory tracking control of the ball.The positioning control contrast experiment of FUZZY-PID controller and VSFOA-PID controller shows,VSFOA-PID controller has smaller steady-state error(the RMSE of FUZZY-PID controller is 2.5108 mm,and the RMSE of VSFOA-PID controller is 1.7658 mm,only 70.33 % of FUZZY-PID controller),shorter adjusting time(the adjusting time of VSFOA-PID controller is 5-12 s,and the adjusting time of FUZZY-PID controller is 6-20 s).The trajectory tracking control experiment of VSFOA-PID controller shows that VSFOA-PID controller has faster tracking speed,the average time to tracking 100* 100 mm2 square is only 8s.