The Path Planning And Simulation Of The Mobile Robot Based On Fuzzy Uncertainty Algorithm

Mobile robot represents the highest achievement of mechatronics, which is oneof the most active fields in the domestic and foreign science and technologydevelopment. Together with the navigation and positioning, obstacle avoidance andtracking, path planning has become the important part of the research of the mobilerobot.This thesis carries out the research on the path planning of the mobile robot. Onthe one hand, planning is considered to be a kind of optimization problem, theenvironment of the robot is represented by a parameter space, so the path planning isto find a best solution in this space. As one of the classic path planning method, theartificial potential method is a kind of modeling method that is only related to systemposition parameters. First of all, this thesis introduces the theory of artificial potentialmethod, and then constructs the gravitation function and repulsion function model.For the local minimum problem that usually occurred, this thesis brings in the newrepulsion field function based on the relative distance to improve the model. Thesimulation results show that the new repulsion function can solve the proposedproblem well; at the same time, it can plan the safe path.On the other hand, path planning can also be seen as a kind of uncertaintybehavior. For the kind of uncertainty behavior that caused by the fuzzy character ofthe things, we take the fuzzy math as the tool to deal with them. The membershipfunction becomes the carrier for fuzzy math to resolve this kind of problem. For thediscussion of the effect of the membership function to controller characteristics, adesign method of membership function based on Gauss-type function is introducedfor fuzzy approximator. By virtue of the adjustable parameters, the “typical model” ofthe membership function is constructed. Take the one-dimensional system andtwo-dimensional one for example, the fact that the approximators are based ontriangle membership function or Gaussian membership function both can approximate the nonlinear system with arbitrary precision is proved to be true. Thesimulation makes comparison analysis of the approximation accuracy and real timeperformance, which provides the precision reference model for the latter design.Based on the above conclusions, this thesis realizes the design of a path planningmethod based on the fuzzy uncertainty algorithm. The main innovation of this methodis that it constructs the two-dimensional fuzzy controller based on the Gaussianmembership function, which makes language quantitative description more inaccordance with human thinking. At the same time, this algorithm has moreadvantage in speeding, so it is more suitable for the real-time situation. Consideringthe complexity of the computation, this thesis makes use of the offline query responsetable, which avoids the large numbers of multi variable matrix computation duringthe process of the fuzzy inference. It also alleviates the computational load of thereal-time inference. The simulation is carried out under the same situation, and theresults show that the fuzzy controller has stable performance; the path based on thefuzzy uncertainty algorithm can avoid obstacles in limited time. The running time isbetter than the results of the traditional artificial potential field method, so it cangreatly enhance the response speed of the robot in the actual environment.