| Mobile robot is a nthetic system which includes multiple functions. The study on the path planning has being an important aspect in the mobile robot research field and turning into a hot spot issue. Its purpose is to find a near-optimization path in the environment where has some obstacles with some optimization indexes. Based on the search of genetic algorithms and particle swarm optimization respectively, the paper finds a good method for path planning.In the part of theory, the general conceptions, characteristics, classify based issue and some familiar methods of path planning are firstly presented. Secondly, this paper puts forward the solutions based on genetic algorithm. Through a variable-length chromosomes coding method, crossover operator and mutation operator's design, smooth and optimization of chromosomes and adaptive parameter adjustment method, the path can be optimized to avoid into local minimum value and get the goal location. Then the standard PSO has been improved. Inertia weight has been changed into linear. Meanwhile the model of the vertexes of obstacles is constructed to describe two-dimensional map of work place of the mobile robot in order to obtain a constrained path function from the start location to the goal location, and further translates the path planning problem into nonlinear constrained function minimum optimization problem. Some invalid particle will be change into random valid particle once again. So extend the search range and keep part from get into local optima. Finally using modified PSO optimizes forcefully the function to get the satisfactory collision-free path solution.In the part of simulation, the influence to the path in which self-adaptive parameters have been adjusted has been discussed. Then simulation and analysis have been carried out in three static environments by two algorithms, and for modified PSO, the dynamic path planning is simulated. The simulation result proved the rational of the algorithms and the method is effective. |
PDF Full Text Request