資源描述:
《基于蟻群算法的機(jī)器人路徑規(guī)劃研究》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、碩士論文基于蟻群算法的機(jī)器人路徑規(guī)劃研究摘要移動(dòng)機(jī)器人路徑規(guī)劃是機(jī)器人研究領(lǐng)域的核心內(nèi)容之一,具有復(fù)雜性、約束性和非線性等特點(diǎn)。蟻群算法(ACA)是最近十幾年發(fā)展起來的仿生優(yōu)化算法,該算法在解決許多復(fù)雜問題方面已經(jīng)展現(xiàn)出其優(yōu)異的性能和巨大的發(fā)展?jié)摿?。本文主要研究靜態(tài)環(huán)境下基于蟻群算法的移動(dòng)機(jī)器人全局路徑規(guī)劃。首先,采用柵格法建立環(huán)境模型,并利用做過改進(jìn)的基本蟻群算法在柵格環(huán)境模型中進(jìn)行路徑規(guī)劃,這些改進(jìn)有:利用偽隨機(jī)比例規(guī)則代替隨機(jī)比例規(guī)則進(jìn)行路徑轉(zhuǎn)移;限制了螞蟻行至當(dāng)前柵格時(shí)下一步允許選擇的柵格范圍;對(duì)啟發(fā)函數(shù)進(jìn)行了重新定義;讓螞蟻根據(jù)轉(zhuǎn)移概率
2、利用“輪盤賭”方法選擇下一個(gè)柵格。其次,針對(duì)基本蟻群算法在某些方面的不足和缺陷提出了三種改進(jìn)算法:針對(duì)螞蟻在搜索路徑過程中落入障礙物陷阱而導(dǎo)致的算法停滯現(xiàn)象,提出了帶夭折策略的蟻群算法;針對(duì)蟻群在路徑搜索初始階段建立的非最優(yōu)路徑上的信息素對(duì)以后蟻群的信息誤導(dǎo)作用,提出了帶獎(jiǎng)罰機(jī)制的蟻群算法;針對(duì)機(jī)器人在實(shí)際工作中的安全避碰問題,提出了基于保守螞蟻的蟻群算法。最后,在蟻群算法的基礎(chǔ)上結(jié)合遺傳算法(GA)提出了兩種改進(jìn)算法:GA.ACA算法和ACA。GA算法,并將其應(yīng)用于機(jī)器人路徑規(guī)劃。為了驗(yàn)證本文所提各種算法的有效性,基于MATLAB7.5軟件開發(fā)
3、環(huán)境設(shè)計(jì)了基于蟻群算法的移動(dòng)機(jī)器人路徑規(guī)劃仿真系統(tǒng)。仿真結(jié)果驗(yàn)證了所提算法的有效性。關(guān)鍵詞:路徑規(guī)劃,柵格,蟻群算法,遺傳算法Abstract碩士論文Thepathplanningformobilerobotsisoneofthecorecontentsofthefiledofroboticsresearchwithcomplex,restrictiveandnonlinearcharacteristics.Theantcolonyalgorithm(ACA)isanewbionicsoptimizationalgorithmdevelopedi
4、nthepastdecade,itshowsexcellentperformanceandgreatpotentialfordevelopmentwhensolvingmanycomplexproblems.ThisthesismainlystudiesglobalpathplanningformobilerobotsbasedonACAinstaticenvironment.Firstly,gridmethodisusedtoestablishtheenvironmentmodelandsomemodificationsalemadetoacc
5、ommodateACAtopathplanningingrid-basedenvironment.Thesemodificationsinclude:usingthepesudorandomproportionalruleinsteadoftherandomproportionalruletochoosepath;limitingthescopeofthenextgridallowedtobechosenbytheants;redefiningtheheuristicfunction;usingtheroulettetochoosethenext
6、鰣dfortheants.Secondly,threeimprovedalgorithmareproposedtoovercomecertaindefects.Inordertoavoidalgorithmstagnationcausedbytheantsfallingintotheobstacletrap,ACAwithabortionstrategyisproposed;Inordertoeliminatethemisleadingofthepheromonethattheantsreleaseininitialstageofpathplan
7、ning,ACAwithrewardandpunishmentmechanismisproposed;Inordertosearchtheoptimalpathwimcollisionavoidance,ACAbasedconservativeantsisproposed.Finally,twoimprovedalgorithmareproposedbasedonACAandgeneticalgorithm(GA):GA-ACAalgorithmandACA-GAalgorithm.Inordertoverifytheeffectivenesso
8、falltheproposedalgorithm,pathplanningsimulationsystemformobilerobots