Research On Sectionalized Controlling Of Motion For Automatic Guided Vehicle

Non-holonomic constraint is characterized by non-integrable differential equations, making the system more difficult to control. On the other hand, non-holonomic constraint exists widely in practical systems, such as automatic guided vehicle (AGV), space robot, and under-actuated underwater ship system and so on, with a great potential for applications in military, industry, civil-application, deep-sea, outer-space and other fields. Since AGV is a typical non-holonomic system, it is of great theoretical and practical importance to study its control issue. Based on the current research status of the AGV, the present work aims to perform in-depth investigation of the motion control problem for tricycle AGV. Specifically, the main tasks of this study include neural dynamic model-based early stage tracking control, energy-efficient strategy-based middle stage tracking control, and last stage tracking and stabilization unified control, as well as the integrated motion sectionalized control (IMSC) in the entire AGV motion process.Based on the systematic research of the current situation around the world, the dissertation systematically analyzes the AGV's structure and category, and several key motion control technologies for AGV's three basic motion forms (path following, trajectory tracking and point stabilization). Finally, several problems needing further investigation are pointed out.By introducing some basic concepts and theorems of differential geometry and non-linear control theories, related knowledge for the non-holonomic system and non-holonomic constraint is given. Besides, a set of math tools for non-holonomic system are generalized, which are used for analyzing the AGV's non-holonomic properties. Taking tricycle AGV as an example, the kinematic and dynamic models for the AGV are developed.According to the hardware structure and software design principle of the AGV system, a reasonable function module partition is proposed and the control architecture of AGV is presented.In view of the initial velocity jump of the traditional tracking controller upon AGV's initial posture error or discontiguous reference trajectory, on the basis of tracking error model, a neural dynamic ideology is introduced and a trajectory tracking controller for AGV, which is based on the biologically inspired neural dynamics model, is proposed. First of all, the proposed tracking controller generates an ideal velocity law using the kinematic controller. Then, the problem of initial velocity jump is solved using the neural dynamics ideology. Finally, the tracking accuracy is further improved using a fast terminal sliding mode controller. Simulation results show that the tracking controller solves the initial velocity jump problem successfully and guarantes the global asymptotic stability of the system.With consideration of the AGV's excessive energy consumption in motion, in light of the thorough analysis of the energy consumption status for the AGV system, an energy consumption model and a kinematic model are developed and a trajectory tracking controller basing on the energy-efficiency strategy is proposed. The core of the proposed controller is an energy optimal control module, which is with the motor energy efficiency function as the objective function. The constraint equations for this module are as follow:the voltage balance equation of the motor armature equivalent circuit and the torque equation are the system state equations, the condition equation of tracking the reference trajectory is the state constraint equation, and a control constraint equation. Finally, the above optimal problem is solved by genetic algorithm and an optimal velocity law is obtained. Simulation results show that the proposed controller achieves trajectory tracking and optimizes the energy consumption.In order to resolve the AGV's tracking and stabilization unified control problem in the final stage, the non-linear model predictive control (NMPC) problem with non-holonomic constraints and control constraints is investigated based on the thorough analysis of the principle and stability of the predictive control method. Also a terminal stability control algorithm based on model predictive terminal control is proposed. In order to improve the tracking accuracy, a state observer is designed to estimate the AGV's state in the dynamic system. Besides, considering the fact that the AGV should have obstacle-avoidance function, an obstacle-avoidance module is designed. The validity of the controller is validated by computer simulation.In order to achieve control with high precision and low energy consumption for the entire AGV tracking process, a method called integrated motion sectionalized control (IMSC) is proposed. In this method, the three forementioned control methods are adopted for the early, middle and last state motion of AGV respectively, according to their characteristics. Through combining the three control methods into one and determining the applicable motion interval for each method by calculation and simulation, IMSC is capable of providing smoothness, energy saving, robustness and global stability for the AGV system.