Executive Petri Net Control System Design For Auto Guided Vehicles

As the robotic industry develops rapidly and the working scenarios become more complex,mobile robots are widely applied in the industrial processes.Robots are required to transport among several target locations following a predefined path under,in most cases,indoor environments.In order to fulfil this need,robots are supposed to percept the accurate positional information and move precisely along the track,as well as owning a stable and reliable execution system to deal with the tasks under multiple application scenarios and mission needs.A Petri-net based control system for automatic navigation vehicles is presented in this paper.The main contents are as follows:1.Design methods of Executive Petri Net(EPN)for auto guided vehicles:including some basic modules of robots,the subnets of pose estimation task and hybrid target tracking task.A new TBB model of functional unit for dealing with the problem of resource conflict in the design of EPN,improving the activity of EPN.Dividing the task of robot into low-coupled functional modules and representing the executive tasks of robots by unified TBB models.Connecting TBBs through external transition to get an EPN according to the execution logics.The executive algorithms of functional unit TBB and EPN,as well as the flow-process diagram of EPN are presented explicitly in this paper.According to the given algorithms and flow-process diagram,the execution program of the robot can be implemented quickly to realise the engineering applications.2.The design of the subnet of pose estimation task for tracking of auto guided vehicle,based Kalman filter and kinematic model of four-wheel robot: applying Kalman filter algorithm to filter the data of wheel encoder,then fusing the data of the encoder and the gyroscope with the aid of the kinematic model of four-wheel robot.In this way,achieving the precise estimation of the robot's position and pose.3.Introduction of the design of hybrid target tracking task subnet.A design of ahybrid switching trajectory tracking control algorithm: the PID coordinate tracking control system is put forward if the distance between vehicle position and target position is higher than the preset threshold,in order to reduce the amount of the control calculation.Using fuzzy predictive controller for the steady tracking of target track if the distance between vehicle position and target position is within the threshold.The hybrid navigation controller has the advantages of simple control,self-tuning parameters and engineering realizable.4.Taking Songling's robot as the experimental platform,STM32F4 as the carrier of robot task execution algorithm,to process data and control the robot's movements.A path of arc plus a line as well as a parking track is designed for the experiment.The EPN is designed especially for the experiment.The correctness and feasibility of the hybrid switching navigation controller and EPN are verified through the experiment.Taking the advantage of the Petri net,the execution process of the task is described clearly in this paper.The control logics and programs are able to be verified by the means of Petri net at the same time,thus benefits the efficiency of the development of program in different tasks.