本文選題：路徑規(guī)劃 + 柵格法環(huán)境建模��； 參考：《蘭州理工大學(xué)》2017年碩士論文

【摘要】：利用移動(dòng)機(jī)器人對(duì)太陽(yáng)能光伏板進(jìn)行清潔的首先任務(wù)是解決路徑規(guī)劃的問(wèn)題。本文研究了在西部地區(qū)大型太陽(yáng)能光伏板發(fā)電站的環(huán)境背景下,移動(dòng)平臺(tái)式清潔機(jī)器人基于改進(jìn)蟻群算法進(jìn)行全局路徑規(guī)劃的解決辦法。在過(guò)去環(huán)境模型中只顯示二維長(zhǎng)度信息的基礎(chǔ)上,進(jìn)行拓展改進(jìn)——增加危險(xiǎn)度這一概念,更加逼近現(xiàn)實(shí)的模擬智能路徑規(guī)劃問(wèn)題;并且實(shí)現(xiàn)對(duì)太陽(yáng)能光伏板發(fā)電站三維地圖信息的環(huán)境建模,最后進(jìn)行全局三維路徑規(guī)劃問(wèn)題的研究與應(yīng)用。主要工作如下:(1)針對(duì)建設(shè)在西部地區(qū)大型太陽(yáng)能發(fā)電站光伏板清潔難題,提出輪式移動(dòng)機(jī)器人清潔的解決方案。以太陽(yáng)能光伏發(fā)電站實(shí)況模擬搭建環(huán)境模型,對(duì)光伏板清潔機(jī)器人的全局路徑規(guī)劃問(wèn)題進(jìn)行了研究。將太陽(yáng)能光伏板設(shè)置為環(huán)境地圖中的障礙物,提出了一種基于概率理論與蟻群算法相結(jié)合的改進(jìn)柵格環(huán)境建模方法,通過(guò)在搜索路徑中選擇經(jīng)過(guò)概率加權(quán)的策略,進(jìn)行全局路徑規(guī)劃的模擬仿真實(shí)驗(yàn)。實(shí)驗(yàn)結(jié)果證明,較傳統(tǒng)建模方法,改進(jìn)后的建模方法具有一定的優(yōu)越性。(2)為了克服蟻群算法中控制參數(shù)難以確定、易陷入局部自由度、早熟等現(xiàn)象,采用云模型理論對(duì)蟻群算法進(jìn)行改進(jìn),通過(guò)對(duì)云隸屬函數(shù)的參數(shù)控制,實(shí)現(xiàn)算法的自適應(yīng)調(diào)整策略,最終減少算法的迭代速度。與基本蟻群算法的對(duì)比實(shí)驗(yàn)結(jié)果證明,改進(jìn)蟻群算法的光伏板清潔機(jī)器人路徑規(guī)劃方法提高了機(jī)器人路徑規(guī)劃效率,且能安全避開(kāi)障礙物。(3)針對(duì)清潔機(jī)器人在工作環(huán)境中,因?yàn)闅夂蜃兓蜕硥m天氣影響,在規(guī)劃路徑中出現(xiàn)的動(dòng)態(tài)障礙物等沖突問(wèn)題,希望提高機(jī)器人路徑規(guī)劃中的越野性能和工作效率,提出一種三維路徑規(guī)劃的方法。依據(jù)地勢(shì)海拔的高低不同,進(jìn)行概率論分布來(lái)設(shè)置不同的通過(guò)優(yōu)先級(jí),將地勢(shì)海拔的高低變化和最短最優(yōu)的路徑相互平衡,從而智能合成一條距離最短且路徑中高坡與低谷地形難度適中的路徑,利用MATLAB軟件設(shè)計(jì)了仿真實(shí)驗(yàn)環(huán)境,驗(yàn)證了該方法的有效性。(4)根據(jù)文中的仿真實(shí)驗(yàn)結(jié)果,結(jié)合實(shí)際中成熟且可高度自由定制的實(shí)驗(yàn)平臺(tái),進(jìn)行了大量的模擬實(shí)驗(yàn),獲得了珍貴的實(shí)驗(yàn)數(shù)據(jù),并且對(duì)此進(jìn)行研究,得出對(duì)比仿真模擬實(shí)驗(yàn)和實(shí)際工作要求的差距,指出不足點(diǎn)和未來(lái)工作的重心方向。
[Abstract]:The first task of cleaning solar photovoltaic panels with mobile robots is to solve the problem of path planning. In this paper, the solution of global path planning based on improved ant colony algorithm for mobile platform cleaning robot in the environment of large solar photovoltaic power station in western region is studied. On the basis of only displaying two-dimensional length information in the environment model in the past, this paper extends the concept of improving-increasing the risk degree, and approaches the practical simulation intelligent path planning problem. The environmental modeling of three-dimensional map information of solar photovoltaic power station is realized. Finally, the global three-dimensional path planning problem is studied and applied. The main work is as follows: (1) aiming at the problem of photovoltaic panel cleaning in large solar power station in the western region, the paper puts forward a solution to the cleaning of wheeled mobile robot. The global path planning problem of photovoltaic panel cleaning robot is studied based on the environment model of solar photovoltaic power station. The solar photovoltaic panel is set as the obstacle in the environmental map. An improved grid environment modeling method based on the combination of probability theory and ant colony algorithm is proposed. The simulation experiment of global path planning is carried out. The experimental results show that compared with the traditional modeling method, the improved modeling method has some advantages. In order to overcome the difficulty of determining the control parameters in ant colony algorithm, it is easy to fall into the local degree of freedom and precocity. The ant colony algorithm is improved by using cloud model theory. By controlling the parameters of the cloud membership function, the adaptive adjustment strategy of the algorithm is realized, and the iterative speed of the algorithm is finally reduced. Compared with the basic ant colony algorithm, the experimental results show that the path planning method of photovoltaic clean robot with improved ant colony algorithm improves the efficiency of robot path planning, and can safely avoid obstacles. Because of the impact of climate change and dust weather, the dynamic obstacle in the path planning is a conflict problem. In order to improve the off-road performance and work efficiency of robot path planning, a three-dimensional path planning method is proposed. According to the elevation of the terrain, the probability distribution is carried out to set different priorities to balance the variation of elevation and the shortest and optimal path. So we can intelligently synthesize a path with the shortest distance and moderate difficulty of high slope and valley terrain in the path. The simulation experiment environment is designed by using MATLAB software, and the validity of the method is verified. (4) according to the simulation results in this paper, the simulation results are given. Combined with the mature and highly customizable experimental platform in practice, a large number of simulation experiments have been carried out, and precious experimental data have been obtained, and the gap between the simulation experiment and the actual work requirements has been drawn. Points out the shortcomings and the direction of the center of gravity of future work.
【學(xué)位授予單位】：蘭州理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP242

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 江威;吳艷蘭;譚樹(shù)東;馬藝文;;一種多發(fā)生元Voronoi圖的柵格生成方法[J];地理與地理信息科學(xué);2015年05期

2 李翠明;龔俊;牛萬(wàn)才;王，

本文編號(hào)：2043895