Research On Methods Of Intelligent Robust Control For Cleaning Mobile Manipulator Of Large Condenser

The control problem of multi-joint manipulator and mobile robot has been paid a great attention for a long time, and many theoretical results have been reported. However, most of research works are focused on the control issues of the traditional manipulator or mobile robot systems. A mobile manipulator is a robotic manipulator mounted on a moving base, with the function of mobile and operation, which makes it superior to the traditional characteristics of mobile robots and manipulators. It not only has the manipulator operational flexibility, but also increases the size of the robot workspace, with almost infinite work space. For the purpose of improving the heat transfer performance of condenser and increasing the efficiency of thermal cycle of turbo-generator unit, we investigate intelligent automated cleaning system for large condenser based on a mobile manipulator. This dissertation focuses on the methods of intelligent robust control for cleaning mobile manipulator of large condenser. Main results and contributions of this dissertation are as follows:Firstly, the principle of the large condenser, the cleaning method and its shortcomings are introduced. Then, the research progress of the mobile manipulator and robust control of robot are generalized. And, the research significance of this dissertation is presented.In this dissertation, design and research of the large condenser cleaning robot is a kind of mobile manipulator system, and the mobile manipulator is composed of a nonholonomic mobile robot and a honolomic manipulator. In chapter 2, the kinematics and dynamics model and its characteristics of a nonholonomic system and a multi-joint manipulator are introduced respectively. Then, aiming at the centralized control strategy and decentralized control strategy of mobile manipulator, the corresponding kinematics and dynamics model are founded.In the centralized control strategy, as sliding mode control is an effective nonlinear feedback control method of uncertain systems, in chapter 3, a sliding controller is designed based on the bound of parameter uncertainties and external disturbances of mobile manipulator system. The advantages of this controller are sliding control does not need precise mathematical model of the controlled object, and that as long as know the change scope of the error of parameters. The control rules of sliding mode control are to change the switching motion constantly, make the state trajectory of the system arrive the sliding mode. In theory, when the state trajectory of the system reaches and slips on the surface, equivalent control can make the state trajectory slide to the sliding surface and keep on it. However, when the uncertainties and external disturbances are unknown, there inherent chattering phenomenon of sliding control can not be made good use of. The existence of chattering may cause instability of the system, and restrain the application of the sliding mode control in a certain extent. For counteracting the defects of sliding mode control, a sliding mode control system for mobile manipulator based on neural network is designed. The neural network is used to compensate the uncertainties and external disturbances. Based on Lyapunov theorem, the structure of sliding mode controller and the learning algorithm of the neural network are designed. So the stability of the system is guaranteed, and the dynamic performance of the system is improved. The simulation results show that the sliding mode control method based on neural network can weaken chattering phenomenon effectively, and has an excellent dynamic characteristics.In Chapter 4, aim at the existing problems in cerabellar model articulation controller (CMAC), by fuzzifying the space division method of cerabellar model articulation controller, a fuzzy CMAC neural network is proposed. Then, using fuzzy CMAC neural network approaches the model of mobile manipulator, and on this basis, an H_∞controller with adaptive mechanism is proposed. The effects of external disturbances and the reconstruction error of fuzzy CMAC neural network are reduced by using the H_∞control strategy. Theoretical analysis shows that the controller can restrict the effects in a designated area, and all signals in closed-loop system are bounded. Finally, in order to validate the effectiveness of the H_∞control strategy based on fuzzy CMAC neural network, a set of experiences which is compared with computed torque control method is drawn. The simulation results show that the proposed control strategy is better than computed torque control method under external disturbances, has better robust performance.In decentralized control strategy, mobile manipulator is divided into two subsystems in Chapter 5, that is, nonholonomic mobile platform subsystems and holonomic manipulator subsystem. Then, considering the kinematics controller of the mobile platform, Lyapunov function of the two subsystems are designed. The couple between the two subsystems is regarded as disturbances, and aim at some unknown parameters and all unknown parameters, corresponding robust adaptive controller are given respectively. According to the Lyapunov stability theory, the overall mobile manipulator system is stable, and the tracking error and adaptive coefficient error is uniformly ultimately bounded (UUB). The stability of the system is ensured and the dynamic performance is improved. Finally, simulation results show that the robust adaptive controller is effective, and has good robust and adaptive capacity.Considering the size, cost and so on, robot system is usually not equipped with speed measuring devices, and obtains speed information only through position feedback. In Chapter 6, a fuzzy adaptive nonlinear robust observer is proposed. The parameters of system uncertainties are approximated by fuzzy logic, and the robust term restrains external disturbances and reconstruction error of fuzzy logic. The observing errors are proven to be UUB by strictly positive real Lyapunov design. Then, on the basis of the designed observer, a fuzzy adaptive output feedback controller is designed. Parameters of fuzzy system are tuned adaptively based on Lyapunov stability theory, the controller can guarantee uncertainty robot tracking desired trajectory stability, and all signals in the closed-loop system are bounded. Finally simulation results show that the method is effective.Aim at the requirements of condenser online efficient cleaning, a new online cleaning project for condenser is presented in chapter 7. According to the condenser structure and layout, high pressure water jet cleaning and chemical cleaning are combined, and cleaning project is realized by two-joint robot arm. The cleaning system structure is introduced, and the architectures of three generations of cleaning robot are analyzed in detail.Finally, the main innovations of the thesis are summarized, and the fields for further research are expected.