| Working as a nexus in the whole process of production, the automatic guided vehicle (AGV) is the key equipment in the modern flexible manufacturing systems. Analysis of the current AGV Applications,it can be found that AGV researcher prefers adopting PLC or single-chip computer as the core control module, and PLC incurs restrict in the system expansion and low performance in complicated schedule coordination algorithm. However the industrial personal computer also own the short slab in working stability aspect. Therefore, a highly stable and scalable AGV controller with flexible configuration has vital significance for domestic AGV application and development.The system design has two parts,the first one is the distributed control strategy based on AGV on-board controller,another is the centralized control strategy based on Supervisory controller.Firstly, this paper analyze the characteristics of flexible manufacturing systems and the basic functional modules including: motor drive, path navigation, mileage positioning, obstacle avoidance protection and logging is determined through the requirement analysis. The Digital Signal Controller (DSC) is selected as the core of AGV control unit to fulfill requirements in two aspects: schedule coordination algorithm performance and power & cost consumption. The design of AVG controller based on MC56F8366 is superior in peripheral interfaces, flexible configuration and versatility. Its navigation function can be carried out on AGVs of different type, including: magnetic sensor guided car, laser radar AGV car or inertia navigation car. The relative motor driving can be realized by PWM wave or communication via CAN bus. Meanwhile, such controller can not only be used in single-AGV system, but also support multi-AGV scheduling of large-scale flexible manufacturing system. The system in this paper contains the magnetic navigation, CAN bus control electrical module and Zigbee wireless communication module, and on this basis, the paper has also designed the settings panel modules, remote wireless monitoring modules for the entire AGV system.The second part of the paper is the software processes design on the system. Dijkstra's algorithm will be used in the AGV path planning to seek out a shortest path for AGV between the start and end. While moving around, AGV will compare collected landmark parameters with storage information to load up the shortest route. The paper has also raised two scheduling method. One is an distributed control method based on DSC for small number of AGV application, AGV used this method replies task request independently and makes a corresponding respond on actual working conditions. The method is relatively complicated in software processes of vehicle controller. The other is centralized control method, which makes the host computer issue commands according to the feedback of status information AGV. Such method is always adopted in more number of AGV application and has a relatively high requirement in communication. The two methods use the same hardware but different software processes, users can get the properly selection according to actual situation.In a multi-AGV system, each AGV needs to seek out an appropriate path between start and end with no confliction and the time is optimal. The paper bring forward an dynamic path planning algorithm based on time window calculation ,considering large calculation and time consumption of static path planning method. The time window algorithm has fully taken the characteristics of AGV embedded applications (ie, small memory space, real-time) into account, which only concerns about the time window overlap on the current point and implement iterative update for overlapping time windows of low priority and realize a no confliction and time optimal path planning. Simulation experiments and practical experiments show that the time window algorithm used on multi-AGV working condition can achieve no conflicts between vehicles, the time optimal path planning function... |
PDF Full Text Request