機器人工業協會(Robotic Industries Association)指出，從2003到2005年間
，北美機器人工業以20%depingjunnianzengchanglvfazhan。zaiqicheshichangxuqiupiruanyijiwaiguochangshangdeyalibuduanzengjiadebeijingxia，zheyichengjiushiruhequdede
？chengbendepubianjiangdishiyuanyinzhiyi。congjiqirendaozhengtaoxitong
，jiagedouzaibuduanjiangdi。jiqirenxingnengdebuduangaijinyeshiqudongzheyiqiangshizengchangdeyuanyinzhiyi。xianzaidejiqirenkeyizhixinghenduojinianqianhaiwufazhixingderenwu。jiqirenhainengzaigengduandeshijianwanchenggengduodeshiqing
，tigonglegenggaodeshengchanlishuiping。 
   最重要的原因則應該是機器人智能的不斷改進。從機器人問世到現在，在決定部件獲取、檢查自身組件就位狀況、檢測錯誤狀態或相關因素上的智能經曆了數個發展等級。這樣的智能大多是基於用來檢測某個特定狀態的特定傳感器而實現的。 
  比如
，機器人會用“光電眼”，通過識別某個芯片管腳、製動器或其它部件是否就緒來判定是否存在並處於正確的方向。接著，這個“光電眼”連接PLC或者直接連到機器人控製器中。在恰當的時間，機器人會核查“光電眼”來確定某個部件就位並處於正確的方向
，然後才會將之摘取或者進行其它操作。 
   使用“光電眼”或huo者zhe類lei似si的de傳chuan感gan器qi是shi一yi個ge簡jian單dan而er可ke靠kao的de途tu徑jing，也ye是shi正zheng確que的de選xuan擇ze。但dan是shi
，自zi動dong化hua並bing不bu總zong是shi那na麼me簡jian單dan。有you各ge種zhong類lei型xing的de部bu件jian需xu要yao處chu理li，因yin此ci區qu分fen部bu件jian的de方fang法fa可ke能neng更geng為wei複fu雜za。 
   buguanshibujianshizhizaochengxu，jiandandechuansongqidouwufashiyong。biru，fangzaiyouhuatuobangechengshucengdexiangguizhongdebujian，yibanjiuyongyujinshubujian。eryongyoufuzajihegouzaodebujian，ruguomeiyouewaidezhuangzhi，jiuhennanzhunquedingwei。 
二維視覺定位 
   長chang期qi以yi來lai，由you於yu無wu法fa找zhao到dao合he適shi的de方fang法fa來lai處chu理li這zhe些xie複fu雜za事shi項xiang，某mou些xie行xing業ye裏li機ji器qi人ren的de使shi用yong總zong是shi無wu法fa擴kuo大da。最zui近jin這zhe種zhong狀zhuang況kuang終zhong於yu有you所suo改gai觀guan了le。其qi中zhong作zuo用yong最zui大da的de技ji術shu就jiu是shi雙shuang維wei視shi覺jiao係xi統tong。請qing登deng錄lu自zi動dong化hua成cheng像xiang聯lian盟meng(Automated Imaging Association)網站獲取關於此技術的更多信息。 
   2Dshijiaoxitongyoubiaozhunxingyeshexiangjigoucheng，yonglaipaishezhaopianyigongjiqirenjinxingfenxibingcongerjuedingchulinaxiebujian。gongyeshijiaoxitonglishibingbuchang
，danxianzaizaijiage、性能和可靠性上已經發展到可以用於一些應用中，而僅僅幾年前這些應用都還是無法實現的。 
   使shi用yong視shi覺jiao係xi統tong可ke以yi協xie助zhu機ji器qi人ren來lai摘zhai取qu放fang置zhi於yu由you標biao準zhun滑hua托tuo板ban隔ge成cheng數shu層ceng的de箱xiang櫃gui中zhong的de部bu件jian。這zhe是shi將jiang部bu件jian從cong一yi個ge車che間jian傳chuan到dao另ling一yi個ge車che間jian甚shen至zhi在zai同tong一yi個ge車che間jian內nei進jin行xing傳chuan送song的de通tong用yong方fang法fa。如ru果guo沒mei有you視shi覺jiao係xi統tong，製zhi造zao商shang就jiu必bi須xu使shi用yong相xiang對dui較jiao貴gui的de成cheng形xing塑su料liao墊dian料liao或huo者zhe通tong過guo其qi它ta方fang法fa來lai準zhun確que摘zhai取qu箱xiang內nei的de部bu件jian。這zhe種zhong能neng置zhi於yu箱xiang櫃gui中zhong的de成cheng形xing塑su料liao墊dian料liao相xiang對dui更geng昂ang貴gui一yi些xie，其qi中zhong光guang是shi模mo具ju的de設she計ji和he製zhi造zao就jiu需xu要yao6萬到10萬美元。 
   2D視shi覺jiao係xi統tong可ke以yi很hen好hao地di取qu代dai成cheng形xing墊dian料liao或huo者zhe其qi它ta更geng昂ang貴gui的de方fang式shi來lai定ding位wei箱xiang櫃gui中zhong的de部bu件jian。不bu過guo直zhi到dao最zui近jin幾ji年nian
，仍reng然ran有you一yi些xie原yuan因yin使shi得de視shi覺jiao係xi統tong的de使shi用yong苦ku難nan重zhong重zhong
，包bao括kuo不bu同tong批pi部bu件jian顏yan色se的de差cha異yi、箱櫃具體狀況的差異和重複使用的隔板上的記號等等。隨著視覺技術的不斷發展
，這些問題現在都可以很好地解決了。現在的2D視覺係統已經可以定位大多數堆疊於箱櫃內部隔板頂層的部件。 
   zhezhongyingyongyibandefangfajiushishiyongyigeanzhuangzaixiangguishangdeshexiangjilaidingweibujian。shexiangjianzhuangzaizugoudegaodu，yincijiqirenkeyizaishexiangjidixiayidongbingjinruxiangguineibu。zaimeiyicengdeqidian，jiqirenduihuiduishexiangjifalaidegaicengbujiandetupianjinxingfenxi
，ranhouzhugedingwei。 
   如(ru)果(guo)需(xu)要(yao)更(geng)精(jing)確(que)的(de)定(ding)位(wei)，也(ye)可(ke)以(yi)將(jiang)攝(she)像(xiang)機(ji)安(an)裝(zhuang)在(zai)機(ji)器(qi)人(ren)上(shang)
，由(you)機(ji)器(qi)人(ren)將(jiang)攝(she)像(xiang)機(ji)移(yi)動(dong)到(dao)一(yi)個(ge)或(huo)一(yi)組(zu)部(bu)件(jian)上(shang)端(duan)並(bing)拍(pai)照(zhao)，然(ran)後(hou)將(jiang)照(zhao)片(pian)傳(chuan)回(hui)給(gei)機(ji)器(qi)人(ren)進(jin)行(xing)分(fen)析(xi)並(bing)定(ding)位(wei)。當(dang)某(mou)一(yi)層(ceng)上(shang)所(suo)有(you)部(bu)件(jian)都(dou)被(bei)取(qu)走(zou)之(zhi)後(hou)，機(ji)器(qi)人(ren)會(hui)取(qu)掉(diao)隔(ge)板(ban)，並(bing)開(kai)始(shi)摘(zhai)取(qu)下(xia)一(yi)層(ceng)的(de)部(bu)件(jian)。 
   buguanshijiangshexiangjiguding，haishianzhuangzaijiqirenshang，suozengjiadechengbendoubishiyongyigetebiededianliaohuozheqitafangfalaidingweibujiandechengbenyaodideduo。shijiaoxitonghainengdailaigenggaodelinghuoxing

，yichulitongyitiaoxianshangdebutongbujian，huozheshiyingbujiandegengti。youyuchengbenshangdeyuanyin，zheyangdezidongdiebanxieduozaijinianqiangenbenwufashixian，erxianzai
，youleshijiaoxitong，jiqirenfangankexingle

，yekeyifudandeqile。 
   視覺係統給機器人使用帶來的另外一個益處，是降低了用來將部件呈送給機器人的傳送器的成本。在使用2D視覺係統之前，很多部件都必須在由托板傳送器傳送的固定托板上定位。即便是簡單的托板傳送器成本也要3萬美金，而且成本還會隨著傳送器傳送距離的增長而提高。有了2D視(shi)覺(jiao)技(ji)術(shu)
，部(bu)件(jian)可(ke)以(yi)在(zai)相(xiang)對(dui)便(bian)宜(yi)的(de)帶(dai)式(shi)傳(chuan)送(song)器(qi)上(shang)進(jin)行(xing)傳(chuan)送(song)。被(bei)操(cao)作(zuo)員(yuan)或(huo)者(zhe)機(ji)器(qi)人(ren)放(fang)到(dao)傳(chuan)送(song)器(qi)上(shang)的(de)部(bu)件(jian)接(jie)著(zhe)會(hui)被(bei)傳(chuan)送(song)給(gei)機(ji)器(qi)人(ren)。當(dang)安(an)裝(zhuang)於(yu)傳(chuan)送(song)器(qi)末(mo)端(duan)的(de)一(yi)個(ge)攝(she)像(xiang)機(ji)檢(jian)測(ce)到(dao)某(mou)個(ge)部(bu)件(jian)已(yi)被(bei)傳(chuan)送(song)到(dao)末(mo)端(duan)時(shi)，傳(chuan)送(song)器(qi)會(hui)馬(ma)上(shang)停(ting)止(zhi)。攝(she)像(xiang)機(ji)還(hai)會(hui)對(dui)部(bu)件(jian)進(jin)行(xing)定(ding)位(wei)，以(yi)協(xie)助(zhu)機(ji)器(qi)人(ren)將(jiang)之(zhi)摘(zhai)取(qu)。 
   和用來在箱櫃中從隔板上取走部件一樣
，視覺係統在傳送帶上也可以協助機器人準確定位並摘取大多數部件。 [page_break] 

   對定位同一平麵放置的部件，2D視shi覺jiao係xi統tong是shi完wan美mei的de，但dan它ta也ye不bu是shi萬wan能neng的de。在zai過guo去qu幾ji年nian裏li
，已yi經jing有you三san維wei視shi覺jiao係xi統tong可ke以yi用yong來lai定ding位wei不bu在zai一yi個ge平ping麵mian上shang的de部bu件jian。比bi如ru
，這zhe些xie部bu件jian可ke能neng是shi彼bi此ci層ceng疊die的de
，但dan也ye可ke能neng隨sui著zhe部bu件jian的de疊die起qi而er兩liang端duan換huan位wei。此ci時shi2D圖像就不能提供足夠的信息來處理這種位移了。 
   一個已經在實際使用中得到證明的簡單方法是

，在2D攝像機的基礎上再結合激光器光帶。在箱櫃頂部安裝2D攝she像xiang機ji對dui部bu件jian進jin行xing大da致zhi定ding位wei
，並bing確que認ren下xia一yi個ge要yao選xuan擇ze的de部bu件jian。然ran後hou在zai機ji器qi人ren上shang安an裝zhuang另ling一yi個ge攝she像xiang機ji，並bing結jie合he激ji光guang器qi。機ji器qi人ren移yi動dong激ji光guang器qi和he攝she像xiang機ji到dao另ling一yi個ge部bu件jian上shang方fang，接jie著zhe
，激ji光guang器qi會hui將jiang附fu加jia準zhun心xin對dui準zhun部bu件jian上shang的de某mou個ge目mu標biao。部bu件jian上shang的de這zhe個ge目mu標biao可ke以yi是shi一yi條tiao邊bian、一個圓或者其它明顯標誌。通過簡單的三角測量，攝像機就可以立體定位部件的位置和方向。 
   最高端的應用是利用3D視覺係統來定位箱櫃中任意方向的部件，但這顯然困難重重，比如，部件有可能彼此混雜在一起，而且，還要避開箱櫃壁。 
觸覺響應 
   盡管視覺係統是機器人智能傳感器最通用的方法，卻不是唯一的方法。還有一種6dusuiyilichuanganqikeyonglaigeijiqirentigongchujiaoxiangying。zaiyaoqiugaojingdudezuzhuangzhong
，lichuanganqijiuhuibeiyonglaiyindaoliduheshidebujianjinxingcharucaozuo
，biru
，jiangzhouzhuangwucharudaomougekongzhong。daiyoulichuanganqidejiqirenhaikeyiyongyugengfuzadezuzhuangrenwu，birujiangzhuangzhicharudaowaike。zhegeguochengxuyaofeichangzhunque，bingyaojingguohenduogebuzhou。jingguobianchengdejiqirenkeyixiangzhenzhengderenyiyang，qianhouyidongzhuangzhi，zhidaomeiyigebuzhoudouwanquanfuhe。 
   zaiduiyigefuzadebujianjinxingmoguangshiyekeyishiyonglichuanganqilaitigongchujiaoxiangying。yiqianzheshiyoutedingdeshebeilaijinxingde
，danshizhexieshebeikenengwufadadaosuoyaoqiudejingquedu。geijiqirentianjia6D力傳感器，然後加上一個磨盤，機器人就可以在方向改變的情況下仍保持一個不變的力度，抵消了重力影響。 
今後的方案 
   智(zhi)能(neng)傳(chuan)感(gan)器(qi)技(ji)術(shu)在(zai)各(ge)種(zhong)領(ling)域(yu)機(ji)器(qi)人(ren)的(de)成(cheng)功(gong)運(yun)用(yong)中(zhong)起(qi)著(zhe)至(zhi)關(guan)重(zhong)要(yao)的(de)作(zuo)用(yong)。隨(sui)著(zhe)智(zhi)能(neng)傳(chuan)感(gan)器(qi)技(ji)術(shu)的(de)繼(ji)續(xu)發(fa)展(zhan)，機(ji)器(qi)人(ren)在(zai)將(jiang)來(lai)將(jiang)擁(yong)有(you)更(geng)多(duo)功(gong)能(neng)。盡(jin)管(guan)在(zai)進(jin)行(xing)極(ji)其(qi)複(fu)雜(za)的(de)區(qu)別(bie)上(shang)沒(mei)有(you)什(shen)麼(me)可(ke)以(yi)替(ti)代(dai)人(ren)腦(nao)這(zhe)台(tai)超(chao)級(ji)電(dian)腦(nao)

，但(dan)過(guo)去(qu)那(na)些(xie)曾(zeng)被(bei)認(ren)為(wei)不(bu)可(ke)能(neng)的(de)應(ying)用(yong)現(xian)在(zai)對(dui)於(yu)智(zhi)能(neng)機(ji)器(qi)人(ren)已(yi)經(jing)是(shi)“家常便飯”了