Study On Behavior-Based Path Planning Of The Mobile Robot In Unknown Environment

Robot for exploration of rumor is one of mobile robots for exploration of unstructured environment. As the forthgoer for the astronaut's landing, it is very important for the moon exploration and research. Therefore, lots of time and money is spent by many developed countries in designing their own robot for exploration. Supported by the national 863 project of China (No. 2003AA735084), the path planning of the mobile robot is studied in this thesis based on autonomous behavior.As matter of fact, the mobile robot for exploration usually moves on unknown surface on the moon, thus, this thesis emphatically studies design of basic behavior and coordination strategy of behaviors based on the behavioral control viewpoint and the fuzzy control method applied. Based on the two traditional navigation methods of arbitration mechanism and fusion mechanism, the transition mechanism is proposed. The simulation results prove the validity of presented algorithm.Finally, the simulation plant based on VRML is designed by using the interface function between MATLAB and VRML. And by the simulation plant, the foundation of the motion planning and control for the mobile robot in 3-D environment is laid. This thesis is organized as follows:Firstly, the background of the problem and the development of the moon exploration in the world are introduced. The relevant literature about the control algorithm of the path planning for mobile robots are summarized, especially the behavior- based control methods.Secondly, on the assumption that the basic configure of mobile robot is based on the Pioneer II and its improvement, the kinematics of the robot is dealt by linear transformation in order to convenience compiling the fuzzy rules. Then the control scheme is designed for the mobile robot based on behavior.Thirdly, the basic behaviors such as the terrain-analyzing behavior, obstacle-avoiding behavior and goal-trending behavior are designed by using the fuzzy control method to avoid the bad real-time performance of the model-based method. The main purposes of the terrain-analyzing behavior is to macro-plan the mobile robot after analyzing the terrain complexity in the current range and avoid plunging into obstacle blocks. The obstacle-avoiding behavior is to prevent the mobile robot from colliding with obstacle. The goal-trending behavior is to lead the robot trending the goal. In addition, the simulation results based on the 2-D map proceeds to prove the validity of the algorithm.Then, the transition mechanism is proposed to harmonize the behavior based on the two traditional navigation methods of arbitration mechanism and fusion mechanism. The arbitration mechanism only chooses a single behavior at a certain time, that is, only one behavior will affect the output of the system. The arbitration mechanism can result in the fast response velocity, but it is easy to obtain ZENO phenomenon which can be found by the oscillations of the translation velocity and rotational velocity. The fusion mechanism endows every behavior a weight coefficient which determines the proportion of every behavior in the total output. The weight coefficients are generally given by the fuzzy rules or weight rules. The fusion mechanism can lighten the ZENO phenomenon but will slow down the response velocity. The transition mechanism firstly combines two behaviors, which easily produce ZENO phenomenon, into a sub-advanced behavior, that is, the goal-trending-terrain-analyzing behavior and goal-trending-obstacle-avoiding behavior. Then the transition mechanism fuses the two sub-advanced behaviors. The transition mechanism can both obtain the fast response velocity and lighten the ZENO phenomenon. A great deal of simulation results are proposed to prove the validity of the transition mechanism. The simple discussion how to lighten the ZENO phenomenon is also presented, from the simulation results we can see that using more detailed fuzzy rules and inputs is an efficiency method.Finally, the simulation plant based on VRML is designed by using the interface function between MATLAB and VRML. And by the simulation plant, the foundation of the motion planning and control for the mobile robot in 3-D environment is laid.