Vision-based Predictive Tracking Control For Mobile Robots

With the development of hardware,software and control technology,mobile robots have been widely used in military,agricultural,industrial,and service fields.Path tracking control is one of the key and difficult technologies in the field of mobile robots.This technology is not limited to mobile robots,but also can be used in cutting systems,CNC cutting,3D printers,ships and aerospace.Therefore,in-depth study of path tracking control technology in mobile robots can not only enrich the theoretical results of control,promote the development of mobile robots,but also provide solutions for more field motion control technologies,satisfying the high requirements of path tracking control technology in various fields,which have great practical significance and application value.By introducing visual sensors to enhance the environment awareness of mobile robots,it can make mobile robots adapted to specific operating environment quickly and efficiently.At the same time,image processing technology is difficult,and it is easy to lose targets and be affected by the light in complex environment,which brings new challenges to the development of path tracking control technology of mobile robot.Aiming at the problems in Eye-in-hand framework and Eye-to-hand framework,this thesis studies the tracking algorithm based on model predictive control for mobile robots,and the simulation and experiment are completed.The main work is as follows:1.In the Eye-in-hand framework,the problem of mobile robot path tracking predictive control is studied by online learning classification adaptive tracking feature.Firstly,an adaptive image processing algorithm based on optical flow tracking,learning and classification is proposed,which makes the features can be accurately and stably identified and tracked,and solves the problem of easy loss of features under complex background.Furthermore,,a predictive tracking controller is designed based on the traced features,and the asymptotic stability of the closed-loop system is analyzed with Lyapunov stability theory.2.In the framework of Eye-to-hand,a new parametric path prediction tracking control method is proposed to solve the problem that the actual speed and the set speed cannot match rapidly during the path tracking process of the mobile robot.Firstly,to improve the speed of fast matching,a new path updating equation is proposed,and the path parameters are introduced into the prediction performance index to improve the tracking accuracy.Furthermore,according to the proposed path parameterization method,an effective set method is used to solve the predictive control problem under the constraint condition,and a circular path simulation is completed to verify the effectiveness of the algorithm.3.Two experimental platforms under Eye-in-hand and Eye-to-hand frameworks are designed to verify the effectiveness of the algorithm.Firstly,the image processing algorithms of online classification learning is programmed to adaptively recognition and tracking features,and the host computer of mobile robot is programmed,which has the functions of tracking effect display,real-time mobile robot operation parameter display,drawing path and so on.Secondly,a wireless communication system between multi-camera fixed camera and mobile robot is completed.The server-side host computer and client host computer are programmed,which has the functions of data communication encoding and decoding,sending robot control command,real-time running status monitoring,drawing error waveform and drawing path,etc.Finally,some experiments are given to verify the effectiveness of the proposed algorithm.