单AGV最优路径规划及其系统开发

发布时间：2018-07-02 18:17

本文选题：AGV + 路径规划　；参考：《河南工业大学》2017年硕士论文

【摘要】：AGV(自动导引运输车)路径规划在AGV的运行过程中起着关键作用,虽然有关AGV路径规划的研究已经有几十年,但由于AGV所处的工作环境具有复杂程度高、避障困难、优化路径难度大的特点,所以仍然有许多值得研究探索的地方。路径规划方法是对路径规划的关键,它在AGV的运行过程中起着重要作用。在长期的路径规划研究历程中,逐渐涌出了众多的路径规划算法,主要分为两大类,一是传统的算法,二是最近逐渐兴起的智能算法。但不管是传统的算法还是智能算法,它们都具有各自的优缺点,运用这些算法所得到的结果并不一定是最理想的,因此,这就需要研究者对路径规划方法做进一步深入的研究。本文在了解有关AGV路径规划问题的现状、方法及目前存在的问题之上,并考虑到AGV在现实工作环境中的运行情况,针对单激光导航AGV路径规划问题进行研究。运用目前AGV路径规划方法规划出的路径普遍存在长度不够理想、圆滑度欠缺的缺点,因此有必要对现存的路径规划方法进行改进。本文主要采用两种路径规划方法,一种为改进的遗传算法,其改进主要有三处,第一处为对初始种群生成方法的改进,第二处为对变异算子的改进,第三处为平滑算子的引入。采用改进遗传算法的原因主要是其具有较好的并行性的特点,可以同时对多个解空间进行搜索。在本文中,在不同障碍物数量及不同障碍物摆放位置的环境下运用改进遗传算法对单激光导航AGV的运行路线进行规划,通过与基本遗传算法得到的结果对比来分析改进算法的性能,从而探究在不同环境下改进遗传算法相对于基本遗传算法对单激光导航AGV进行路径规划的优劣。针对遗传算法中随机生成的初始种群个体适应度不高的缺点,本文采用另一种AGV路径规划方法,即遗传蚁群算法。在此方法中,初始种群个体不是随机产生,而是来自于蚁群算法每一代的迭代结果,然后经过选择、交叉、变异、平滑得出最终的结果,并经过与其它算法对比,证明了此算法具有一定的改进性。本文针对以上两种不同的路径规划算法,编制相应的MATLAB代码进行仿真验证,并基于C#和MATLAB软件编制成相应的路径规划原型系统,并在两个激光导航AGV经常工作的环境中运用此原型系统进行规划,证明了此原型系统具有一定的实用价值。
[Abstract]:AGV (automatic guided Transport vehicle) path planning plays a key role in the operation of AGV. Although the research on AGV path planning has been for decades, it is difficult to avoid obstacles because of the complexity of the working environment of AGV. It is difficult to optimize the path, so there are still many places worth exploring. Path planning is the key to path planning, and it plays an important role in the operation of AGV. In the long course of path planning research, there are many path planning algorithms, mainly divided into two categories, one is the traditional algorithm, the other is the intelligent algorithm that has been emerging recently. However, both traditional algorithms and intelligent algorithms have their own advantages and disadvantages, and the results obtained by using these algorithms are not necessarily the best. Therefore, it is necessary for researchers to further study the path planning methods. On the basis of understanding the current situation, methods and existing problems of AGV path planning, and considering the running situation of AGV in real working environment, this paper studies the path planning problem of AGV for single laser navigation. It is necessary to improve the existing path planning method because the path planning method of AGV is not ideal in length and lack in roundness. This paper mainly adopts two kinds of path planning methods, one is the improved genetic algorithm, there are three main improvements, the first is the improvement of the initial population generation method, the second is the improvement of the mutation operator, the third is the introduction of the smoothing operator. The main reason for adopting the improved genetic algorithm is that it has good parallelism and can be searched for multiple solution spaces at the same time. In this paper, an improved genetic algorithm is used to plan the running route of a single laser navigation AGV in the environment of different number of obstacles and different location of obstacles. The performance of the improved genetic algorithm is analyzed by comparing with the results obtained by the basic genetic algorithm, and the advantages and disadvantages of the improved genetic algorithm compared with the basic genetic algorithm in the path planning of the AGV for single laser navigation are explored in different environments. In this paper, another AGV path planning method, genetic ant colony algorithm, is used to solve the problem that the individual fitness of the initial population generated by random genetic algorithm is not high. In this method, the initial population individuals are not randomly generated, but the iterative results from each generation of ant colony algorithm are obtained, then the final results are obtained by selection, crossover, mutation, smoothing, and compared with other algorithms. It is proved that the algorithm is improved to some extent. According to the above two different path planning algorithms, the corresponding MATLAB codes are compiled for simulation verification, and the corresponding path planning prototype system is developed based on C # and MATLAB software. The prototype system is used to plan in the environment of two laser navigation AGVs, and it is proved that the prototype system has certain practical value.
【学位授予单位】：河南工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH221;TP18

【参考文献】