Mobile Robot Control System Design Based On Petri Nets

With the rapid development of the robot trade,robots are designed to work in a more complex environment,and their tasks becomes more diverse.Accordingly,how to guarantee their control system dependable and capable to quickly response to variations of their environments and tasks are important issues.Furthermore,it is also valuable to develop the techniques to short the development cycle of robots.To address the above-mentioned issues,we propose a formal design method of a mobile robot control system based on Petri nets(PN).First,task elements are defined according to their functions and required software and hardware resources.Then,they are modeled by a special PN,which is called the task building block(TBB).Each TBB has unified internal structures and interfaces and is mapped to a subprogram to perform the task corresponding to this TBB.Second,according to TBB's 2 inputs and 3 output interfaces,6 logical relations are defined between TBBs.Based on these relations,the method is presented to design global transitions to TBBs.According to a given total job,TBBs are selected,and global transitions are designed to connects them together.As a result,the execute Petri net(EPN)is obtained.Third,an algorithm is designed to execute the EPN,and to call the functions corresponding to the marked TBBs.Hence,the executive logics of the control system can be strictly guaranteed by the formal structure of the EPN.Finally,we developed a mobile robot that utilizes a TX2 chip as the computing platform,and an ARM chip as the control one,respectively.This robot is equipped with a brushless DC motor,four sets of ultrasonic sensors,a laser radar,and a binocular visual camera.Its control system is designed the proposed method.It can perform wireless communication,obstacle avoidance,environmental detection,target recognition,and positioning.Therefore,the feasibility of the proposed method is verified.By the proposed method,we model the robot control system as a PN.With the advantages of graphic intuition and logical rigor in PNs,we can express the task logics clearly and strictly,and ensure the reliability of the control system by the PN analysis theories.Consequently,logic errors can be detected and eliminated during the early stage of design processes.This can effectively reduce the debugging workload.Moreover,by introducing TBB,the functional subroutines are packaged,which facilitates the repeated use of codes.Once the tasks of the robot are changed,the development work can be concentrated on the modification and adjustment of the EPN,which can effectively improve the efficiency of the control system synthesis.