Design Of Tracking Controller Of AGV For Moving Target

Automated Guided Vehicle(AGV)is an automatic transportation vehicle without driver. With a certain amount of weight, it can travel between the departure and the destination. AGV is the key equipment of automatic material flow system and flexible manufacturing system. Today, AGV is widely used in various fields such as machinery industry, automatic storehouse, and material dispatch center and so on in developed countries. With the development of economy in China, some enterprises also start to apply AGV now. According to the navigation methods, AGV can be classified into different types. Among them, the magnetic inductive guiding method is mainly used for the most of commercial AGV. Such as high cost and long period of the long navigation path building and Path maintenance, difficultly modification are the shortcomings of the navigation method. JinLin University Intelligent Vehicle Group has developed commercial AGV which navigates based on computer vision; this new type of AGV has the best navigation flexibility and higher artificial intelligence. It can go along with the white reflectors, and it can recognize some symbols on the white line, such as acceleration, deceleration, and can stop at any working procedure we want. In auto assembly line, some work piece need to go ahead and follow the work piece which was suspended by the catenaries, so the two parts can be circulated in the same velocity. This form is beneficial to operating workers to assemble. This is a tracking move target question but not a navigation method through computer vision.. This is the aim of the dissertation and this is the necessary process to perfect Assemblage AGV. All research, design and experiment were based on the Assemblage AGV manufactured by IV Group. IV Group has designed some single controller to the AGV, and proved that the control effect is good in some condition but the single controller can not work good effect under all circumstance. Fast response in the biggish err and stabilization in small err is the question, so multi-controller is the method to overcome the conflict. For commercial AGV, accuracy, smooth and steady quality of navigation and fast adjusting quality to the deviation are the significant targets of its performance. The design of controller is the key to guarantee it, and this is the dissertation research substance. According to the content, this paper is made up of five parts. 1. Kinematics model of AGV, identification the model of the drive machine system and the sensor system model building. 2. Design of the optimal tracking controller for moving target. 3. Design of the fuzzy tracking controller for moving target. 4. The fusion of optimal controller and fuzzy controller. 5. Test the fusion controller by experiment. First part, we build a simple kinematics model according to the AGV's structure features and relation between the AGV and the path. Static feature and dynamic feature of driving system is analyzed. We get the transfer function by applying M sequence signals as input response signals, driving system output data is identified through least square means. Because AGV's system is very complicated, so the process of calculate is simplified in some extent. It is proved that the influence of the result after simplifying and actual effect is unanimous basically, so we can design the control system in simplifying models. The sensor occupies very important position in the control system, its precision often influences the precision of the whole control system directly so we have chosen the German SICK Company two-dimensional position sensor DMP of the high-accuracy high-stability and derive the measure models by measuring the income data. In section of the optimal controller design, transverse deviation, lengthways deviation and pointing deviation have been set up as the vectorial combination in the state space. This is very important to design the optimal controller. Linear Quadratic model is chosen according as the system's characteristic and course, confirming the best feedback gain matrix after simulation and experimentation on AGV. We use a changeless gain matrix in the whole process, three deviations can...