National Kaohsiung First University of Science and Technology

FirstTech Institutional Repository
Not a member yet
    18584 research outputs found

    Development of Measurement System for On-line Tool Wear

    No full text
    [[abstract]]在銑削加工過程中,加工刀具會與工件和碎屑產生摩擦與碰撞,從而使刀具鍍層逐漸磨損。當刀具磨耗過大且鍍層過薄時,刀具的移除材料能力將隨之降低,從而造成加工品質的變異。因此,為減少加工品質變異,若能檢視甚至量化刀具磨耗程度,將有助於刀具可用程度的掌握,穩定刀具的移除材料能力,減少加工品質變異,進而提升加工效益。本研究開發一可線上量測刀具磨耗之刀具微影像檢視系統(ToolMicro-imageInspectionSystem,TMIS),包含刀具微影像機(ToolMicro-imageRobot,TMIR)與刀具檢視模組(ToolInspectionModule,TIM)等。其中,刀具微影像機可自動擷取CNC加工機上的不同刀徑的刀具磨耗影像,而刀具檢視模組則從磨耗影像中計算出刀具狀態,包含刀具缺陷與磨耗量。而此系統更可即時將刀具狀態透過WIFI傳送至刀具磨耗估測系統,從而提供線上刀具狀態。在研究成果上,所開發的刀具微影像檢視系統可檢視直徑介於6-25mm的刀具,根據刀具不同的旋轉角度,可自動識別刀具結構與刀具磨耗重點區域,解析度達0.03mm,磨耗量測精確度達95%。因此,本系統將可提供CNC機台線上刀具的影像、磨耗量、與可用程度的即時資訊,精進加工過程的穩定性。[[abstract]]Inthemillingprocess,thecuttercoatingisgraduallywornduetofrictionandcollisionamongtheworkpiece,chops,andcutter.Whenthetoolwearistoolargeandthecoatingistoothin,thetoolremovalcapacitywouldbereducedandresultedinqualityvariation.Hence,toreducethequalityvariation,weneedtoevaluatethedegreeoftoolwearforimprovingthecontroloftoolavailability,stabilizingthestabilityofthematerialremoverate,reducingthequalityvariability,therebyenhancingprocessingefficiency.Thisresearchpresentsatoolmicro-imageinspectionsystem(TMIS)foron-linemeasuringtoolwearincludingatoolmicro-imagerobot(TMIR)andtoolinspectionmodule(TIM).TheTMIRcanautomaticallycapturethetoolwearimageswithvarioustooldiameters.TheTIMcanderivethetoolstatusfromtheimages,includingthetooldefectsandwearsizes.Inaddition,thissystemcantimelytransfertoolinformationtothetoolwearestimationsystemviaWIFIforprovidingtheon-linetoolstate.Intheresearchresults,theTMISisabletomeasurethetooldiametersrangedfrom6to25mm.Accordingthetoolangles,thissystemcanidentifythetoolstructureandwearfocusareawithresolution0.02mmandaccuracy95%.Hence,theproposedsystemisabletoprovidetheCNCmachinetoolwiththeon-linetoolimages,wearsizes,andavailabilitytoimprovemachiningstability

    Analysis and Optimization of Traffic on Highway with Greenshield Model

    No full text
    [[abstract]]台灣運輸網絡發達。然而高峰時段高速公路的交通通常由於缺乏足夠的設施來引導司機或提供交通信息,導致交通堵塞。為了提供更好的服務質量,需要更多的先進設備和傳感器來收集更多的資料,如速度,交通流量和進一步分析和處理能力,作為智能交通系統的信息。本論文通過收集高速公路各個部分的交通數據提供旅行時間和交通量估算,來研究使用Greenshield的模型。傳感器設備包括環路感應線圈、車輛檢測卡、微處理器、Ethernet光電轉換器等,可以收集建立不同Greenshield型號的即時數據。我們可以在各種條件下獲得所需的參數,從而可以通過Greenshield的模型的速度-密度關係和流量達到最佳的吞吐量。本論文提出一種旅行時間估計演算法。利用每個路段環路車輛偵測器收集每分鐘車流、車速數據作為旅行時間估算依據。我們透過傳輸系統將這些數據傳送到交通管理中心數據庫進一步應用在Greenshield’s模型建立速度、密度與自由速度之間的關係。最後將這些現場數據和Greenshield’s模型取得一個適當權重因數進行旅行時間估算。並透過模擬結果顯示基於即時演算法所得到的Greenshield’s模型是有效的,可提供在快速公路(Expressway)上進行準確的旅行時間估算外,還能連動智慧型運輸系統(IntelligentTransportationSystem),提供即時(Real-Time)資訊增進運輸系統的安全、效率及舒適性,減少交通對環境的衝擊。[[abstract]]TransportnetworkinTaiwaniswelldeveloped.Howeverthetrafficinhighwayduringrushhoursoftenresultincongestionduetothelackofenoughfacilitiestoguidedriversorprovidetrafficinformationinadvanced.Forprovidingbetterqualityofservicemoreadvanceddevicesandsensorsarerequiredtocollectmoredatesuchasspeed,trafficflow,andcapacityforfurtheranalysisandprocessastheinformationinintelligenttransportationsystem.ThisthesisinvestigatestheuseofGreenshield'smodelbycollectingtrafficdatainvarioussectionsoftheExpresswayforprovidingthetraveltimeandtrafficestimation.Sensordevicesincludeloopcoil,vehicledetectioncard,microprocessorandEthernetphotoelectricconverter,whichcancollectrealtimedataforestablishingdifferentGreenshield'smodels.WecanobtaintherequiredparametersinvariousconditionssuchthattheoptimumthroughputcanbereachedthroughtheGreenshield'smodel'svelocity-densityrelationshipandflow.Inthisthesis,areal-timealgorithmfortraveltimeestimationisproposed.Webuildaloopvehicledetectoroneachsectiontocollectdatasuchastrafficflow,speedandsoonastraveltimeestimates.WetransferthesedatathroughthetransportsystemtotheTrafficManagementCenterdatabaseforfurtherapplicationintheGreenshield'smodeltoestablishtherelationshipbetweenspeed,densityandfreespeed.Finally,thesefielddataandGreenshield'smodelareobtainedwithanappropriateweightingfactorfortraveltimeestimates.AndthroughthesimulationresultsshowthattheGreenshield'smodelbasedonthereal-timealgorithmiseffective,canprovideaccuratetraveltimeestimatesonExpressway,butalsolinkedtotheIntelligentTransportationSystem,providingReal-TimeinformationtoenhancethesafetyefficiencyofthetransportsystemandComfort,reducetheimpactoftrafficontheenvironment

    Combination of Symmetry in Thermal and Texture in Visible Imagery for Pedestrian Detection

    No full text
    [[abstract]]行人偵測被廣泛應用於智慧交通系統或監控等多種不同領域,一般而言,基於可見光感測器於日間光線充足情況下,可提供有效之影像資訊,然於光線不足環境中則無法運作;熱影像感測器於夜間環境中,因無環境背景熱源干擾,可獲得顯著之行人資訊,然而,當行人與環境熱源相近時,通常為日間環境中,則效能會因而遞減,基於上述,本研究目的主要在提出一個結合熱影像與可見光之全天候行人偵測演算法。本研究分別獨立對熱像與可見光感測器訓練出一組人形分類模型,並應用於後續之行人偵測,上述兩者之差異主要基於熱像與可見光影像特性之不同,進而分別採用不同之特徵,並結合支援向量機(SupportVectorMachine)作為人形分類之依據。於熱像部分主要透過所提出之對稱權重HOG(SW-HOG:Symmetry-WeightedHOG)作為人形外觀特徵,其主要概念為整合HOG特徵於水平與垂直方式之對稱關係,以提升HOG特徵之有效鑑別度;可見光部分則以透過HOG及GoogLenet卷積式類神經網路(ConvolutionNeuralNetwork)最後一層所獲得之表示式分別獲得影像中紋理與RGB色彩空間資訊,最後透過決策融合方式達到整合不同影像資訊之目的。於實驗部分透過所建置之行人資料庫,其中包含四段靜態背景與兩段動態影像序列作為驗證演算法之依據,此外,並同時實作文獻中常用之行人偵測演算法,與所提出之演算法進行分析比較,於所獲得之FPPI曲線圖可發現所提出之演算法可提高文獻中行人偵測之準確度。[[abstract]]Pedestriandetectioniswidelyusedinintelligenttransportationandsurveillancesystems.Ingeneral,thevisiblesensoriseffectiveindaytime,butitfailsinthepoorlightingsituation,suchasnighttime.Onthecontrary,thethermalsensorcanwellsensetheheatofthepedestrianbodywellinnighttime.But,itgenerallyfailsincaseofthatthetemperatureofbackgroundisclosetopedestrianbody,especially,indaytime.Accordingly,theobjectiveofthisstudyistoproposeapedestriandetectionalgorithmbyfusingvisibleandthermalsensorssothatitissuitableforbothdaytimeandnighttime.Sincethevisibleandthermalsensorshavedifferentcharacteristics,themodelsforvisibleandthermalsensorsaretrainedindependentlybasedondifferentfeatureswithSupportVectorMachine(SVM).Forthermalsensor,aneffectivefeaturecalledSymmetry-WeightedHOG(SW-HOG)featureisproposedtoimprovethediscriminabilityofthetraditionalhistogramoforientedgradients(HOG)feature.Forvisiblesensor,thecombinationofHOGfeatureandtheoneextractingfromGoogLenetConvolutionNeuralNetwork(CNN)isusedtomodelthetexturepropertiesofthepedestrianappearance.Afterthat,thefusionmechanismisproposedtofusethedetectionresultsfromtwosensors.Inexperiment,sixvideos,fourfromstationaryplatformandtwosequencesfrommovingplatform,areusedtovalidatetheproposedalgorithm.Besides,severalalgorithmsintheliteratureareimplementedforcomparison.BasedontheFPPIanalysis,themethodproposedinthisstudyoutperformstheother

    The Improved Study of Singing-voice Separation Technique Based on Robust Principal Component Analysis

    No full text
    [[abstract]]歌聲對於歌曲裡擁有許多重要的資訊,例如:歌手聲音、歌詞、情感以及歌曲主旋律等等。近年來對於歌聲分離的相關語音研究已經趨近成熟且多元,但離完美的分離,仍然還有一段距離。在本論文中,我們提出了讓穩健型主成分分析能夠針對不同歌曲調適出此方法的最大效益參數,同時觀察此分析方法在分離上會有的缺陷問題並針對其提出改善方法。在實驗部分,我們使用了MIR-1K資料庫中有男有女、歌聲與伴奏比例大小不同的歌曲來驗證改善後的性能,數據結果顯示我們所提出調適方法與改善方法都能讓其分離技術的效果擁有提升。[[abstract]]Asingingvoicehasalotofusefulinformationforasong,suchas:singervoice,lyrics,emotionsandsongtheme,etc.Inrecentyears,researchonsinging-voiceseparationhasbecomematureandmultifaceted,butfromtheperfectseparation,thereisstillsomedistance.Inthisthesis,weproposetoapplyrobustprincipalcomponentanalysisandadjustthegainparameterforsongswithdifferentsingingtoaccompanimentratio,Wealsoobservethedefectsofthisanalysisandimproveit.Intheexperimentalpart,weusedtheMIR-1Kdatabasewithdifferentgenderandsingingtoaccompanimentratiotoverifytheeffectivenessofimprovedmethod,Theresultsshowthatproposedmethodcanimprovetherobustprincipalcomponentanalysis

    Implementation of Android Malware Detector Based on Permission and Dynamic Analysis

    No full text
    [[abstract]]智慧型裝置所帶來的便利性加快了成長速度,促使在行動網路上有了明顯的發展,但在便利性的誘因下,往往埋藏了許多危機,近年來不停出現各種智慧型手機的攻擊,導致受害者資訊洩漏以及財力上的損失,政府逐漸重視智慧型手機日益普及的發展與保護,積極推動手機應用程式檢測制度,來加速規劃防範潛在資安威脅。本論文研究的主題是針對Android開放式系統進行安全性研究,實現第三方權限系統與汙點追蹤分析,將其整合並強化之。其中,在汙點追蹤技術上為了達到更好的檢測之有效性,是透過現有已知的Web漏洞進行滲透分析,來改善用戶端資料在雲端上暴露的疑慮,因而有效阻止惡意應用程式的行為,藉由權限來企圖存取手機資訊與個人隱私。[[abstract]]Theconvenienceofsmartdeviceshasacceleratedthepaceofgrowthandledtoasignificantdevelopmentonthemobilenetwork.However,undertheconvenience,manycrisesareburied.Inrecentyears,therehavebeenvariousattackstosmartphonessothatthevictimsleakageinformationandfinanciallosses.TheGovernmentgraduallypayattentiontothegrowingpopularityofsmartphonesandprotectionandactivelypromotethemobileapplicationtestingsystemtoacceleratetheplanningtopreventpotentialsecuritythreats.ThethesisstudiesthesecurityofAndroidopensystem,realizestheanalysisofthirdpartyauthoritysystemandstaintracing;thenintegratesandstrengthenstheircapabilities.Inordertoachievethebetterdetectioneffect,weproposeusingexistedWeb’sweaknessforinfiltrationanalysiswhichcandispeltheuser'sdoubtstoexposedinformationonthecloud.Thereforeitcaneffectivelypreventthemaliciousbehaviorbytheapplication,employingthepermissiontoaccessmobilephoneinformationandpersonalprivacy

    Design of Energy Management System with Electric Power Dispatching and Optimization Management Mechanism

    No full text
    [[abstract]]本文提出一個具電力調度與最佳化管理機制之能源管理系統(EnergyManagementSystem,EMS),有別於以往文獻針對負載做卸載與最佳發電排程動作,本文以時間區段內選擇利潤正值為優先考量,不用因為負載需求過大而停止運轉動作,對於時間電價的買賣選擇,提供數據給使用者參考電力供給的最佳利益情況。本文首先以分散式能源系統輸出資料整合,建立成資料庫,透過系統連結輸出,藉由EMS智慧調度,依照一年四季不同需求情況做模擬,搭配台電的電力買賣相關費用,機動調控系統連結到電網端之電力買賣交易值,使供給電力及出售額外電力達最大值,同時,能源管理系統調度介面上,協調系統發電不平均之影響,優先考量儲能裝置內備用電力狀態,再執行賺取利潤結果。最後,能源管理系統依據發電資料解析,於哪一時段相對的供需價格差額最大,依供給與負載需求的結果,分析及調整能源供給及儲能端數據,並做電力調度說明。[[abstract]]Inthisthesis,weproposeanenergymanagementsystem(EMS)withelectricpowerdispatchingandoptimizationmanagementmechanismthat’sdifferentfromotherreferenceswhichusingload-sheddingandoptimalpowerschedulingmethod.Ourproposedsystemprioritizesthepositivevalueofprofitwithinaperiodoftimeandprovidestheoptimizeddataofpowerusagetotheusersforhavingthemaximumprofit.Ontheotherhand,thesystemwillstillcontinueitsoperationevenwhenoverloaded.Firstly,EMScollectsdatafromthetargeteddistributedenergysystemandcreatesadatabaseafterintegratingthedata.DataiscollectedfromthedatabaseandprocessEMSdispatchingaccordingtothesimulatedresultofdifferentseasonconditionsinoneyearandthecorrespondingpowerpriceofTaiwanPowerCompany.Then,thetradingamountwiththepowergridisregulatedtoensuretheuser’smaximumprofit.OntheEMSdispatchinginterface,whileadjustingtheimbalancestateofpowergenerated,thetradingamountwiththepowergridisregulatedbasedontheconditionofbatterystoragesystemThen,theamountofpowergeneratedfromeveryhour,everydayandeverymonthareanalyzedbyEMStodeterminewhichperiodoftimehavethelargestpricedifferencebetweenthesupplyanddemand.Basedontheresultafteranalyzingsupplyanddemand,EMSadjuststhedataofthesupplyandbatterystoragesystemthenmakesanannotationoftheanalyzeddata

    All-Digital Pulse-Expansion-Based CMOS Digital-to-Time Converter

    No full text
    [[abstract]]本篇論文提出以脈衝擴增方式實現的數位至時間轉換器(Digital-to-TimeConverter,DTC),無需採取游標卡尺或是數位至類比轉換器等複雜且高成本的方式,其所實現之電路架構較為精簡與節省成本。本研究採用全數位架構來設計與實作,與以往的全客製設計方法較為不同的是,全數位架構是以可程式方法來調整非均質元件並聯組上的均質元件數量來控制脈衝改變量,所以較為簡易省時,並且進一步提出全數位脈衝混合架構來實現脈衝擴增,以有效改善解析度。本研發電路由脈衝產生器(PulseGenerator,PG)、脈衝擴增電路(PulseExpandingCircuit,PEC)以及時間扣抵電路(TimeSubtractor,TS)三大區塊所組成。PG用來產生一個固定時間寬度的脈波,為脈衝擴增進行準備;本電路的主軸電路:PEC包含了脈衝擴增單元(PulseExpandingUnit,PEU)和多工器,其功能為可程式調整脈衝擴增量的大小進而控制時間擴增寬度;最後TS可將PG所產生的固定寬度脈波移除,如此可使最後之時間輸出只和PEC有關。本論文實作一個四位元的數位至時間轉換器,以TSMC0.35-μm2P4M製程實作,其面積僅僅只有0.045mm2。6顆晶片經由量測後,時間解析度平均為16ps,積分非線性誤差約為±0.4LSB,平均功率在每秒一百萬次取樣下為0.2mW。本次所提出的電路不需要先進製程,也可以達到低成本、低功耗且非常良好的解析度。[[abstract]]InsteadofusingVernierprincipleoracostly,complicateddigital-to-analogconverter,thispaperpresentsanall-digitalCMOSdigital-to-timeconverter(DTC)byusingpulse-expandingmechanism.Inthisstudy,theexpandingresolutioniscontrolledbyutilizingall-digitalcircuit,inotherwords,adjustingthenumberofhomogeneousgatesbyprogrammingmethodtoimplementpulseexpandingratherthanadjustingthesizeratiothatcaneffectivelyreducethesemiconductor-designingtime.Moreover,pulseexpandingisachievedbypulse-mixingschemethatcansignificantlyenhancethetimingresolutionandenabletheDTCtobeconcise.TheDTCcomprisesapulsegenerator(PG)forgeneratingapulse,apulseexpandingcircuit(PEC)forprogramingtiminggeneration,andatimesubtractorforremovingthetimewidthofthepulse.ThePECcomprisesonlyadelaychaincomposedofproposedpulse-expandingunitsandamultiplexer.Foraccuracyenhancement,apulseneutralizationtechniqueispresentedtoeliminateundesirablepulsevariation.A4-bitDTCisfabricatedina0.35-?mTSMCCMOSprocessandithasasmallareaofnearly0.045mm2.Sixchipsweretested,allofwhichexhibitedanimprovedresolution(approximately16ps)andlowintegralnonlinearity(lessthan0.4LSB).Thepowerconsumptionwas0.2mWwhenthesampleratewas1Msamples/sandthevoltagesupplywas3.3V.TheproposedDTCnotonlyhasfavorablecostandpowerbutalsoachievesanacceptableresolutionwithoutrequiringanadvancedCMOSprocess

    Miniaturization of transmission line with arbitrary length and its applications to compact Wilkinson power dividers

    No full text
    [[abstract]]本文提出一任意電氣長度的傳輸線小型化設計方法,將其應用在威爾京生分波器,設計的操作頻率為1.0GHz,在中心頻率時保有傳統型威爾京生功率分配器的特性。本文提出一種縮小傳統型電路的方法,將任意電氣長度、任意阻抗之傳輸線,以ABCD矩陣推導出方程式,並將此傳輸線等效為T型傳輸線,並以馬蹄形耦合線等效此T型傳輸線,此架構是由兩段左右對稱馬蹄耦合線並聯一開路殘段所組成,其中該開路殘段可配合電路的需求設計成一y=jb的落地元件,可有效地增加電路的靈活度。經過相關理論分析以及本文整理的設計流程得到此架構之各項參數,再進行模擬及實作。小型化威爾京生分波器經過理論的分析,決定小型化架構中的電氣長度θ=12°,透過公式得到各項參數後,模擬並實作電路,最終電路的量測結果為反射損耗S_11=-27.65dB、S_22及S_33=-31.694dB,插入損耗S_21及S_31=-3.387dB,兩輸出端之隔離度S_32=-29.591dB,頻寬定義為|S_11|?15dB時FBW=54.57%,電路內面積為351.65?mm?^2,縮小比例為95%。本文提出的小型化威爾京生分波器不僅縮小了面積,在未來也可透過並聯一導納y=jb的方式以取代原本的開路殘段。透過本文的設計流程能方便實現於各種PCB電路板上,達到易設計、成本低廉,電路佈局簡易等優勢。[[abstract]]Adesignofminiaturizationoftransmissionlinewitharbitrarylengthinthisarticle.And,itapplytoWilkinsonpowerdividers.Wilkinsonpowerdividersarewidelyusedtosplitthepowerinmanycommunicationsystemsduetothelowinsertionlossandgreatlyhighisolationbetweenthetwooutputports[1].Thequarterwave-wavelengthtransmissionlinesarenecessarypartsoftraditionalWilkinsonpowerdividers,buttheyalsooccupythelargecircuitarea.Thisthesisispresentedamethodbymeansofreplacingquarter-wavelengthtransmissionlinebytheT-networktransmissionline.Accordingtoreference[2],thespurline’smatrixcanbelisted.And,lettheT-networktransmissionlineequivalenttothespurline.Thestructureiscomposedoffoursetsofspurlinesandopen-endedstubs.Theopen-endedstubcanmodifyitintoatransmissionlinewiththesamefeaturebutwithdifferenttype.Acompactpowerdividerthatoccupies95%areaofthatatraditionalpowerdividerisdesigned.Basedona15dBreturn-losscriterion,themeasuredFBWis54.57%.At1.0GHz,theinsertionlossesare-3.38dB,theinputreturnlossis-27.65dBandtheisolationislowerthan-25dB.Themainadvantagesofthisminiaturizedtechniqueareitscompactsize,simpledesignprocedure,lowcostandnotcomplexcircuitlayout

    Manufacturing flow time estimation: case study of IC back-end assembly

    No full text
    [[abstract]]現今的產品製程越來越複雜,尤其半導體業更是如此,且在製造過程中所需要的原料和零件也是日益增加,而使用的材料又會根據產品需求有所不同。在IC封裝製程中有許多的材料都是需要先解凍才能使用的,解凍耗時數小時甚至是需要一天的時間,而解凍後的使用期限最多只有2天,因此材料很容易就會過期。對於產線人員來說,要決定解凍的時間、材料種類及數量是ㄧ大挑戰。本研究將透過迴歸樹並根據分析IC封裝測試廠的實際資料,發展預測流程時間的模型,以幫助產線人員計算產品預計到達的時間,決定該在什麼時候解凍材料避免造成浪費。本研究的目標為當產品進入封裝製程時即可預測到達Molding的時間,建立的模型除了可以直接從各大站預測到達Molding的時間外,也設計另一種分段式的預測模型提供產線人員使用,例如:在進入Rivet站時,以當下的的製程情況計算出預計到達WireBonding的時間,以及從WireBonding到Molding的時間,預測出流程時間後,除了可以幫助Molding進行備料之外,也可以幫助WireBonding準備材料。最後本研究針對不同的站別,分別建立各站可使用的預測模型,模型主要是根據在製品庫存、各站當時使用的機台數目、訂單需求量,以及距離交期的剩餘天數來預測流程時間,不論是直接預測或是分段式預測到達Molding的時間,都能夠達到相同的預測準確度,且研究結果也顯示,當產品進入Rivet進行預測後,接下來進入各個大站時應該更新現場的情況,並使用當站的預測模型重新計算流程時間,以提升預測的準確性。[[abstract]]Themanufacturingprocessesbecomemoreandmorecomplicatednowadayssuchassemiconductormanufacturing.Thematerialsandcomponentsusedintheprocessesalsoincreaseandtheywouldbedifferentaccordingtowhichproducttypesareused.Therearelotsofmaterialsneedtotakehoursorevenonedaytodefrost,theexpireddateisshortandeasilyoverdue.Itisabigchallengeforproductionlinetodecidewhentothawoutthemultiplematerials.Thisstudyaimstodevelopthemodelsforpredictingthemanufacturingflowtimetosuggestproductionlinewhentodefrostthematerialsandavoidtowastetoomuchmaterials.TheregressiontreeapproachisusedandanalyzedthehistoricaldatafromanICback-endassemblycompany.ThetargetistopredictthemanufacturingtimetoMoldingstagewhenaproductmovesintotheassemblyprocess.ThemodelcandirectlypredicttheflowtimefromeachstagetoMoldingstage,andthereisalsoanothermodelcanforecasttheflowtimesegmented.Forexample,whenaproductmoveintoRivetstage,themodelcanevaluatetheprocesstimefromRivettoWireBondingandWireBondingtoMoldingbasedontheprocessstatusatthemoment.ItcanhelpnotonlyMoldingtopreparethematerialbutalsoWireBondingstage.Finally,theresearchdevelopsthedifferentmodelsforeachstages,andthemodelsestimatethemanufacturingflowtimeaccordingtoworkinprocess(WIP),machinenumbersineachstage,orderquantitiesandtheslacktimetoduedate.Theforecastaccuracyforpredictingdirectlyissimilartousethesegmentedmodels.TheresultshowsafterpredictingtheflowtimetoMoldinginRivetstage,whentheproductmovesintonextmajorstage,itshouldupdatetheprocessstatusandre-predicttheflowtime,anditcanimprovetheforecastaccuracy

    A Study on the Correlativity of Traffic Accident Factors in Kaohsiung

    No full text
    [[abstract]]根據內政部警政署民國一百零五年高雄地區道路交通事故的統計資料顯示,肇事原因主要為駕駛過失(其次為車輛機件故障及行人不慎),其中以未依規定讓車、未保持行車安全距離、違反號誌管制與指揮、左轉彎未依規定、未保持行車安全間隔及違反特定標誌(線)禁制等前六項統計數量最多的肇事原因,由此六項肇事原因可推論事故發生以人為駕駛過失為主要原因,此研究是透過研究肇事因素,並分析出有哪些環境因素是容易發生車禍的,並以引為誡。本研究以探討車禍肇事之「交通肇事因素相關性之研究以高雄地區為例」為題,了解肇事原因、肇事者年紀、肇事路口型態以及天氣之間的相關性,並以機車統計,因為機車數量比汽車還多,而機車所發生之車禍事故比汽車都要來得嚴重,所以此分析以機車為主,資料來源為交通大隊95年至105年之肇事資料,將資料分成五種範圍,全部範圍統計、第一當事人機車和第二當事人機車第一及第二當事人分開各自統計、第一當事人機車和第二當事人汽車以第一當事人機車統計、第一當事人汽車和第二當事人機車以第二當事人機車統計,利用SPSS統計軟體進行主成份分析,分析結果顯示各項因素間多為不相關。騎乘機車因包覆性較低,因此在駕駛過程中車速過快摔車,或是被其他車輛撞擊,所造成之傷害會比駕駛汽車來得嚴重,因為在機車並沒有像汽車一樣有安全帶和安全氣囊,唯一的保護只有頭部的安全帽而已,所以造成傷害的面積相當大.騎乘機車注意安全,務必眼觀四方、耳聽八方。[[abstract]]NegligentDrivingisthemainreasonforthecausesoftrafficaccident(followupbyvehicledefectsandrecklesspedestrian)accordingtothestatisticaldataoftrafficaccidentinKaohsiungin2016fromNationalPoliceAgency,MinistryoftheInteriorRepublicofChina(Taiwan).Recklessdriving,Tailgating,Runningredlights,ImproperTurns,Tightsidedistance,andRunningstopsignsarethetop6causesoftrafficaccidentsandthebestproofsofhumannegligenceisthemainreasonforthecausesoftrafficaccident.ThisStudyisbasedonthecausesoftrafficaccidenttoanalyzewhatkindsofenvironmentalfactorscouldresultintrafficaccident.ThisisastudyontheCorrelativityofTrafficAccidentFactorsinKaohsiung.Thestudywillbaseonunderstandingcorrelativityofthecausesoftrafficaccident,perpetrators’age,accidentsurroundingsandweather.Tobeclarifying,thestudyisbasedonthestatisticsdatafrommotorcycle,asthenumbersofmotorcyclesarewaymorethancars,andusuallycausedseriousaccident.Thestudysourcescomefromtheaccidentreportsfrom2006-2016fromTrafficPoliceBrigade.Thereportswillbeviewedin5scopes:Fullscope;Firstparty(motorcycle)andsecondparty(motorcycle)accidentbasedonfirstparty;Firstparty(motorcycle)andsecondparty(motorcycle)accidentbasedonsecondparty;Firstparty(car)andsecondparty(motorcycle)basedonfirstparty(motorcycle);Firstparty(car)andsecondparty(motorcycle)basedonfirstparty(motorcycle)basedonsecondparty(motorcycle).ByusingSPSSstatistic,theresultshowsthefactorsareirrelevant.Ridingamotorcycleismoredangerousbecauseofthelowcoveragetoaperson.Thereisnoseatbeltorsafetyairbagformotorcyclebutonlyhelmet.Thereforethedamagesofridingamotorcyclesuchasfallofforhitagainstcarareusuallymoreseriousthandrivingacar.Itisveryimportantformotorcycledriverstopaymoreattentionontheroadandkeeptheireyesandearsopen

    0

    full texts

    18,584

    metadata records
    Updated in last 30 days.
    FirstTech Institutional Repository
    Access Repository Dashboard
    Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇