The Research Of The Servo Control Method About The Flexible Manipulator

1. Proposition of the subjectTraditional industrial robot arms are generally all designed intothe heavy-duty structure in order to avoid producing orientation errorand causing any vibration which is not hoping. Though the heavy-dutystructure of this kind of rigidity is of great benefit to the accuratelocalization of the robot , the ratio of load to quality is quite small andthe speed is low. In recent years, the light-duty flexible connecting rodmanipulator has already aroused the attention of different fields due tothe constant improvement of manufacturing industry`s demand andthe need of space application. Compared with the rigidity connectingrod manipulator, the flexible connecting rod manipulator has verylight structure and high load--self weight ratio. So its energyconsumption is lower, operation space is larger, production efficiencyis very high and its response is fast and accurate. At present it canovercome many disadvantages of industry's manipulator such asstructure ponderosity,low efficiency,small operation space and slowresponding speed etc., it is a development trend of the next generationrobot, and have important research meanings.It is difficult to get the accurate mathematics model because theflexible manipulator turned out to be a kind of nonlinear time-varietysystem of highly complicated and highly coupling. therefore, themodeling of the flexible manipulator and problem of controllinghave been a problem more difficult to solve all the time . This paperhave carried on research and analyse to modeling and control of theflexible manipulator .2. Difficult point of the subjectThough flexible mechanical arms have a series of advantages than hardmechanical arm, light structure and long size cause mechanical arms have goodflexibility. This make the control and modeling of the flexible mechanical armbecome questions with challenge. The main reason is as follows. 1. The flexible arm has high pole point of vibration and lowdamping, but it is prone to cause flexibility distortion and shake in thecontrol process. 2. The flexible arm is an infinite distribution parameter systemwhich generally adopt low steps system approximate because of thecomputer,imprecise trait of sensor and system noise. The system isapt to bring the control overflow and measurement overflow. Theseoverflows can maybe sometime cause systematic unstability undersome conditions. 3. Driver and sensor are not configuration with same locationamong the flexible arms, cause system input and output causalitypresent minimum phase place characteristic under the influence ofelasticity . 4. Flexible arm is a kind of non-linear time-variety system withhighly complicated and highly coupling. Its non-linear couplingincludes the flexible generalized coordinate and coupling of the jointangle as well as joint angle and its derivative.3. Overall design of the flexible manipulator system In order to discuss the question of control of the flexible manipulator moretrully , more concretly, the structure has been designed which is showed in Fig.1in this paper. This flexible manipulator system is composed of three part: control;execution and sensor.4.The research of the control model Flexible manipulator is a non-linear time-variety system with thehighly complicated and highly coupling., so it is difficult to getaccurate mathematics model. At present the mostly flexiblemanipulator systems adopt mechanism modeling, and it can not getthe accurate mathematics model for some complicated actual systemsto adopt traditional mechanism method such as theorem of physics ,chemistry ,etc. which are proved by a large number of projectpractices. Even though the mathematics model is built upconstrainedly, it is a differential equation group with high-order andnon-linear. Because we neglected a lot of factors and supposed a greatdeal of preconditions in the modeling process, this made this modellose the initial modeling meaning. So it has important actual meaningto study the control methods of flexible manipulator on the basis ofsystem distinguish . The part of Takagi-Sugeno fuzzy model rule are fuzzy variables and anotherpart are the input and output linear function. This text utilizes this trait ofTakagi-Sugeno fuzzy model to linearize the nonlinear system partially and carryon the research about broad sense and predict control method on this basis. Fuzzy modeling can be divided into the structure identification and parameteridentification like the other modeling methods. The structure distinguish includesthat which kind of fuzzy models are adopted ,confirm input and output variable,confirm the fuzzy set and subjection function corresponding to input and outputvariable , that is to say , partition of input and output space, confirm regular figureof the fuzzy model ,etc.. The parameter distinguishes includes adjustment to theparameter of the fuzzy model according to a certain criterion. Generally speaking,structure distinguish and parameter distinguish is a unitary part, it can get thefuzzy model which accord the precision by replacing the two distinguishnumerous and repeatedly . This text utilizes the separated distinguish method of premise partand conclusion part of T-S fuzzy model method. According to systeminput and output data, this text go on the fuzzy cluster first to inputvariable, confirming subjection degree function distribute of premisepart input variable , and then utilize least square method to distinguishthe parameter of the conclusion part under this kind of division. Andthe verification of the model is carried on by the other input andoutput data.5.The research of the control algorithm In the actual application, the sort of control algorithm is a lot.Because traditional PID control algorithm has some advantages suchas simple , good robust and high dependability ect.. it has been one ofthe most wide-ranging control algorithms in the industrial productionprocess , especially apply to set up the certainty control system of theaccurate mathematics model. But for these non-linear, time-varietyuncertainty and the severity delay systems, it is difficult to establishthe exact mathematics model, so the element of quality of traditionalPID control algorithm is not very good. PID control algorithm regarderror as one of basic regulation items, and the differential functiononly works when the system has appeared the obvious deviation, andthat belongs to a afterwards control method, therefore, the PID controlalgorithm can't well suppress systematic overshoot in control effect. Classical feedback control theory established on the basis oftransmitting function and modern control theory based on thefoundation of state space description have a great defect that theaccurate mathematics model of controlled objects is needed to know.All kinds of minimum variance controller generally need to know thetime delay of the target, the control precision will reduce greatly if thetime delay is not accurate; The pole point assignment self calibrationcontroller is very susceptible to the system stairs degree. This highdemand of model precision limits the application of the selfcalibration control algorithm in complicated industrial control process.Flexible manipulator is a kind of non-linear time-variety system withthe highly complicated and highly coupling, so we must look for onekind self-adaptive control algorithms that demand the lowmathematics model and high robustness . Generality predictive control is a kind of prediction control...