Font Size: a A A

The Research Of Contour Following Error Control And Compensating Methods For Multi-axes Motion

Posted on:2005-04-14Degree:DoctorType:Dissertation
Country:ChinaCandidate:B X XiaoFull Text:PDF
GTID:1118360152455950Subject:Power electronics and electric drive
Abstract/Summary:PDF Full Text Request
By way of the core technique for robot, intelligent control technique has an important action, and it will have an important theory and practical meaning to research it. The paper analyzes the following error and contour error in multi-axes motion path control for robot based-on the math model of bi-axial system, and puts forward correlative compensating methods. Compared with traditional methods, those methods don't change the structure of former system, and put a feed-forward or coupling compensating module on the control loop of each axis, so it can overcome the disadvantage of unforeseeable problems in multi-axes motion, and may obtain a certain effect.The main research contents are following:1. Firstly, the research background is introduced in this paper, and universal manipulator is used as object, also the controller design and alteration experiences are introduced. This chapter uses its system structure and joint motor math model and control mode for reference, and presents the math model of single joint finally.2. The chapter 2 discusses the relation between servo gain and contour error for motion path system based-on simplifying position servo system math model. By answering differential equation, the thesis analyzes how the dynamical characteristics of the motion control system influence the contour error based-on the model of bi-axial system in detail. Finally some useful conclusions are obtained.3. In motion control system, the differences of the dynamic characteristics of each axis are the main reason of the contour error. The chapter 3 will design a high-powered feed forward compensator for each axis to improve the dynamic characteristics and then to make different axis characteristic match. Because of the physical difficulty of getting the high rank differential signals of input, the compound control is hard to realize. To solve this problem we use two different kinds of Neural Networks to identify the contrary control object and use each Neural Network as feed forward module. The first network is BP network with one hidden layer, and the second network is linear status Neural Network based on linear system dynamic equation.4. The Research of fast Neural Network training algorithm is taken attention at all times. By analyzing and research, We know the GAs(Genetic Algorithm) is a kind of global & parallel & random searching optimize algorithm and has a good performance on global searching; and that the BP algorithm does quite well in local searching. By adding BP algorithm into GAs we have a new algorithm called Hybrid Genetic Algorithm (HGA) used as the training algorithm oftwo Neural Networks, it may integrate their advantages and learn from other's strong points to offset one's weakness. Also in this paper, its principle and strategies are analyzed and summarized, finally we compare it with BP algorithm, as a result it has the advantage of fast convergence and it will not converge at local nadir through simulation.5. The chapter 4 discusses the structure & characteristic & learning method of fuzzy neural network chiefly. The fuzzy neural network system, which is combination of the fuzzy system and neural network, is one of the modern control strategies that have already brought broad attention. The primary characteristics of fuzzy neural network are that it can identify fuzzy logic rules automatically and can adjust membership function and correlative weight values, and each parameter in fuzzy neural network has clear physics meaning.6. It is another method and will be reasonable for multi-axes motion to adopt cross coupling control to compensate and reduce contour error. The chapter 5 proposes a new Cross-coupling compensation method for contour error based on fuzzy neural network. It also explains principle, algorithm and realization of this compensator. Through the adaptive ability of fuzzy neural network, this method can dynamically provide each axis with additional compensating information according to system contour error, and then make the following error and control action of each...
Keywords/Search Tags:multi-axes motion control, robot, servo gain, servo system dynamic characteristic, following error, contour error, compound control, neural network, training algorithm, genetic algorithm, contrary model, fuzzy neural network, cross-coupling control
PDF Full Text Request
Related items