| Ball&Plate system is the two-dimensional extension of Ball&Beam system in space plane. It is a controlled object with strong coupling, complexity, nonlinear, uncertain parameters, mechanical friction and external random factors, and compared with Ball&Beam system, the controlling Ball&Plate system is more complex, the characteristics of the nonlinear and instability performance more prominent, the usability and classic of experimental study make Ball&Plate system become an important research platform in control theory research gradually. In addition, the Ball&Plate system is widely mentioned in the analysis and modeling of nonlinear systems, computer vision detection, motion control, mechanical design platform, multi-target trajectory planning, and many other fields in various fields such binding characteristics of the system Study challenging. Therefore, this paper selects Ball&Plate system as the research object.Shuffled frog leaping algorithm(SFLA) belongs to the swarm intelligence algorithm, after the algorithm was put forward, it was applied to control engineering by many scholars, and there have been many excellent engineering application cases. Also, SFLA can be taken to do offline tuning and optimazation of control parameters, this kind of bounding "offline" control algorithm is widely used in the simulation experiment, the oneness of control parameters restricts its more widely application. If it combines SFLA with the control algorithm by ―online‖ integration, in the control process of system, realizes online auto-tuning and online self-tuning of control parameters, and improves system’s control quality, this kind of control strategy has no case applied to the Ball&Plate control study. This paper takes Ball&Plate system as the research object, does research and exploration about its control strategy, by introducing SFLA to do PID control parameters’ auto-tuning and online self tuning, finally achieve optimal control about Ball&Plate system.In this paper, it takes Ball&Plate system which is based on visual inspection as the object, and studies on the control of the Ball&Plate system. First of all, this paper analyzes the physical structure and characteristic of system, then, sets up linear mathmetic model of the system. Secondly, it optimizes Ball&Plate system’s visual detection scheme, and improves the system’s anti-jamming capability. Thirdly, it puts forward a SFLA-PID control algorithm that combines SFLA with the traditional PID control method by ―online integration‖, and then it successfully applies the control algorithm to the design of Ball&Plate system SFLA-PID controller, and realizes the Ball&Plate system’s motion control. The experiments’ results show that the SFLA-PID control algorithm proposed in this paper is feasible, the SFLA-PID control algorithm is convnient and effective, and it has good robustness. It is successfully applied SFLA-PID control algorithm to Ball&Plate system’s motion control, the control precision and control accuracy in fixed-point motion are high, and the fixed trajectory motion control has good effect. In the process of the system control, it realizes PID control parameters’ online auto-tuning and online self-tuning, drops the impact of uncertainty and disturbance to control accuracy as much as possible, improves the control effect, and completes control objectives quickly.The main content and innovation of this paper are described as follows:Firstly, it expounds the background, significance and research status of the real study object and theoretical research object by the numbers, analyzes the physical structure and characteristic of system, then uses Lagrange mechanics principle to establish dynamics model of the system, chooses the angle control volume to do system linear, and sets up the mathematical model of the system after complete linearization.Secondly, in view of the lack of visual detection method in Ball&Plate system, this paper does improvement and optimization in the image preprocessing step, to improve accuracy of ball centric localization and the overall performance of control system, and designs the following visual inspection optimized method: image smoothing, edge detection, OTSU threshold selection binary, centric extraction and ball positioning optimization scheme. Physical experiments’ results show that, using optimized visual detection method, extraction centric results and ball positioning results are more accurate, and makes the system has better anti-jamming capability.Thirdly, it utilizes that SFLA’s thought of elimination, evolution, inheritance, puts forward a SFLA-PID control algorithm that combines SFLA with the traditional PID control algorithm by ―online integration‖, designs SFLA-PID controller, and achieves Ball&Plate system’s fixed-point motion control. The results show that the SFLA-PID control algorithm is feasible, fixed-point motion control has good effect, the speed of optimizing parameters is fast, also it has the adaptability of parameter online auto-tuning and self-tuning, ultimately, it make the system quickly achieve stability.Fourthly, it selects the SFLA-PID control algorithm, PID control algorithm, Fuzzy-control algorithm to do compared experiments about fixes-point motion, taking into account that an experiment is repeated many times, not every time it can achieve the setting control requirements or control precision, and puts forward a new evaluation index—control accuracy. After several times and a variety of real-life experiments, it finally analyze and compare results from the control precision, control time and control accuracy, it finds that: SFLA-PID control algorithm can not only achieve the setting point control requirement, it has the highest control precision and control accuracy, the control effect is best; Then it uses SFLA-PID control algorithm to do single-cycle triangle trajectory motion control, tracking speed is fast and control effect is better.Finally, this paper summarizes the work content, puts forward the existing problems and follow-up studies in the research. |
PDF Full Text Request