資源描述：
《移動(dòng)機(jī)器人的路徑規(guī)劃與路徑識(shí)別研究》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(kù)。
1、蠶豢瓣菠大學(xué)硪童磷突生學(xué)位論文繁l燹摘要隨著科舉技術(shù)的進(jìn)步和機(jī)器入學(xué)的發(fā)展，移渤機(jī)器人的戚用越來(lái)越廣泛。作為移動(dòng)機(jī)器人“智能”的重要糊索一一路徑規(guī)劃與路徑識(shí)別，成為具露理論意義秘實(shí)用價(jià)值的藿要課題。尤其當(dāng)機(jī)器人硬《牛系統(tǒng)的糖度在短期內(nèi)不麓褥妥解決瓣請(qǐng)流下，對(duì)算法懿磷究藏顯得蘢為鬟要。對(duì)于路徑規(guī)劃問(wèn)題，本文首先在藻本遺傳算法的基礎(chǔ)上，掇出了一種帶稿發(fā)式的遺傳算法，在簡(jiǎn)單環(huán)境中規(guī)劃_出一條最優(yōu)路徑。其次針對(duì)復(fù)雜環(huán)境中豹磅經(jīng)援劃閹題提出了禁忌一遺簧冀法，矮遺傳算法提供莠牙援索主攝黎，嵌入禁繇搜索的個(gè)體竄行搜索方姣
2、，使算法散能褥戳改避。仿真表明，該方法能有效地提高算法的收斂速度與解的質(zhì)量。最后根據(jù)機(jī)器人工作的安際情況，提出了多約束條件下的路徑規(guī)劃問(wèn)題，通過(guò)全局規(guī)劃垮局部規(guī)劃相綴合瓣方法，技蜀了一條滿足無(wú)瑾、輟逶攜箍不超避羧度蠹簸簸滔徑。目霹，在所有仿真研究之后，從理論上對(duì)各種算法的收斂性進(jìn)行了分析。對(duì)于路徑識(shí)別問(wèn)題，本文采用BP神經(jīng)網(wǎng)絡(luò)的研究方法。在附加動(dòng)量項(xiàng)法期自適應(yīng)學(xué)習(xí)速率法的蹙礎(chǔ)上，提嫩了分階段學(xué)習(xí)法。著以籀二屆全國(guó)大舉生凝囂太嘏筏大賽靜場(chǎng)德為辯象逡行儔寞，實(shí)駿裘韜，致遂靜辜聿經(jīng)霞絡(luò)擎羽收斂速度快、能跳離塌部極
3、小值，髓識(shí)別率高，具有較強(qiáng)的抗干擾和抗噪能力。關(guān)鍵詞：移動(dòng)楓器入路徑蕊捌遺傳算法禁忌搜索路徑識(shí)剮BP神經(jīng)網(wǎng)絡(luò)囂南鶼技大學(xué)碩士耢究生學(xué)位論文第l{燹AbstractWiththedevelopmentofscienceandtechnologyaswellastherobotics，robotshavemoreandmoreapplicationsinvariousfields。Thepathplanningandthepathrecognition，asimportantfactorsinrobotinte
4、lligence，haveimportantvalueintheoryandpractice．Whentheprecisionprobleminrobothardwaresystemcan’tbesolved，thealgorithmresearchappearsparticularlysignificant．Astopath—planningproblem．thispaperfirstputsforwardageneticalgorithmwithheuristictoplananoptimalpathf
5、orasimpleenvironment．Thenaimingatthepathplanningincomplexenvironment，tabu-geneticalgorithmhasbeenpresented，whichadoptsgeneticalgorithmtosupplythemainflameofparallelsearchandembedstheindividualserialsearchmodeoftabusearch，thusthealgorithmcanbedeveloped．Thes
6、imulationresultsindicatethatthemethodcanefficientlyimprovethisalgorithminconvergencespeedandsolutionquality．Furthermore。accordingtotherobot’Srealworkcondition，thepaperpresentsapath—planningproblemwithmulti—restrictions，Bycombiningglobalplanningandlocalplan
7、ning，theshortestpaththatsatisfiescollisionfreeandtransitgoodsnon—excesscanbefound．Finally,afterallthesimulationresearch，thepaperhasanalyzedvariousalgorithmsconvergenceonthetheory+Astopath—recognitionproblem，thepaperusestheBPneuralnetwork．Onthebasesofaccess
8、ionalmomentummethodandself-adaptivestudyspeedmethod，asubsectionstudymethodhasbeenproposed．Andthenthesimulationmakesthegroundofthesecondnationalcollegestudents’robotTVcompetitionasanobject。Simulationresultssho
