Study On Path Planning And High Precision Track Control For Unconstrained Mobile

A lot of research works on trajectory tracking or path planning bymeans of adaptive control, fuzzy logic, AI technique, fuzzy-neuroapproach, neural network, etc, have been reported, but most of them usearm robots or mobile robots as control objects. Arm robots usually haveseveral movement restrictions because of mechanical structures, and cannot work freely without a command. Mobile robots usually have thecapability to recognize their surroundings and can act autonomously byusing many sensors equipped with.On the other hand, an unconstrained object such as a ball on a plate isable to move freely and has no ability to recognize the environment. Sothe unconstrained object can not control its behavior by itself. All of thesefactors make trajectory tracking of the unconstrained object difficult.Ball and plate system is a typical object for trajectory tracking orpath-planning research of unconstrained mobile. It is a mechanical system,which controls a ball's position or moving state on a plate throughdynamically adjusting the plate's rotary angles. As the extension oftraditional ball and beam system, ball and plate system is used as astandard benchmark to inspect diverse control schemes because of itscharacteristic of multi-variable, strong coupling, parameters uncertaintyand nonlinear.The task of study on path planning and high precision track controlfor unconstrained mobile is based on the theory of visual servoing. It isplanned to achieve the following goals on the ball and plate equipmentBPVS-JLUⅠwhich is independently developed by our laboratory.Firstly, building the system's mathematic model by its dynamics andkinematics analysis. Secondly, stabilizing the ball at a point and control itrolling from one point to another on the plate. Thirdly, putting forwardsome control schemes to make the ball track given trajectory. Lastly,searching an optimization path for the ball in complex environment, forexample, some obstacles put on the plate, and control the ball track thepath. The dynamic look-and-move visual servoing structure whichcontains double closed loops is applied for the equipment BPVS-JLUⅠ.The mathematic model of ball and plate system is the basic foranglicizing its kinematics and control characteristic. It is difficult to buildthe system's mathematic model by Newton mechanics laws, because theball and plate make up of complex movement when the ball rolls on theplate. Euler-Lagrange dynamics equation is more suitable to analysis thiskind of system which contains strong coupling, because it needs only theenergy of the system. This paper introduces the ball and plate equipmentBPVS-JLUⅠ, builds its mathematic model by Euler-Lagrange dynamicsequation, and also give out the simple description of the system. Thesimulation model is built in Matlab, and some simulation andexperimental results are completed. The results show that it is accurate ofthe dynamics analysis and the mathematic model.The plate is drived by stepping driving system in BPVS-JLUⅠ. Theroutine stepping driving system is a kind of open loop system, so it can'tconquer the errors from information transfer, drivers and some othernonlinear factors. In order to implement high precision position control, aclosed loop control method for stepping driving system is presented inthis paper. A closed loop controller with a forward feedback controller isapplied in a new scheme for stepping driving system in BPVS-JLUⅠ. Theexperimental results show that this scheme can achieve high precisionposition and velocity control in inner loop.Fuzzy control is most broadly applied in the field of intelligentcontrol. Some fuzzy control schemes are studied in the process ofdesigning the visual servoing controller. In the conventional IF-THENfuzzy inference model, all the input items of systems are usually put intothe antecedent part of each fuzzy rule. Consequently, both the numbers offuzzy rules and parameters increase exponentially, and defining fuzzyrules becomes extremely difficult for large-scale systems. In this paper,SIRMs dynamically connected fuzzy inference model is proposed for theball's track control. For each input item, a dynamic importance degree isgiven and a SIRM is constructed. The dynamic importance degree isdefined as the sum of a base value insuring the role of the input itemthrough a control process, and a dynamic value varying with controlsituations to control the influence of the input item. This paper getssatisfactory results in the experiments of the ball's stability control,moving control and trajectory tracking control. Compared with results ofother research organization, the ball's velocity and tracking precision areall improved.The path-planning problem in ball and plate system is also discussedin this paper. A new path planning method based on genetic algorithms isput forward to find an optimization way for the ball in complexenvironment. It firstly builds a map for the ball's moving space by gridmethod. Then a string of grids expressed by sequence number are orderlyarranged as a path individual. Many individuals constitute the searchpopulation. At last, genetic operators are acted on the population, and theoptimization path is selected according to the standard of the shortest path.Because the path is discontinuous, this paper uses an interpolationapproach named movement track method to construct a slick continuouspath for the ball. The path planning method is proved to be very effectiveby experiments.In all, this paper do a lot of academic and experimental research onball and plate system, and basically achieves its given goals. Because ofthe typification of ball and plate system in its dynamics and kinematicsanalysis and control issues, the work of this paper will be contributiveto the following research on similar dynamic systems.