Uma ubuhlakani bokwenziwa buthathwa njengohambo olusuka ku-A ukuya ku-B, isevisi yekhompuyutha yamafu iyisikhumulo sezindiza noma isiteshi sesitimela esinesivinini esikhulu, kanti i-edge computing yitekisi noma ibhayisikili elihlanganyelwe. I-Edge computing iseduze nohlangothi lwabantu, izinto, noma imithombo yedatha. Yamukela inkundla evulekile ehlanganisa isitoreji, ukubala, ukufinyelela kwenethiwekhi, namandla ayisisekelo ohlelo lokusebenza ukuhlinzeka ngamasevisi kubasebenzisi abaseduze. Uma kuqhathaniswa nezinsiza zekhompuyutha zamafu ezisetshenziswe phakathi nendawo, i-edge computing ixazulula izinkinga ezifana nokubambezeleka okude kanye nethrafikhi yokuhlangana okuphezulu, okuhlinzeka ngokusekela okungcono kwesikhathi sangempela kanye nezinsizakalo ezidinga umkhawulokudonsa.
Umlilo we-ChatGPT uqalise igagasi elisha lokuthuthukiswa kwe-AI, okusheshisa ukucwila kwe-AI ezindaweni eziningi zokufaka izicelo ezifana nezimboni, ukudayisa, amakhaya ahlakaniphile, amadolobha ahlakaniphile, njll. Inani elikhulu ledatha lidinga ukugcinwa futhi lifakwe ikhompuyutha ukuphela kwesicelo, futhi ukuthembela efwini kuphela akusakwazi ukuhlangabezana nesidingo sangempela, i-edge computing ithuthukisa ikhilomitha lokugcina lezinhlelo zokusebenza ze-AI. Ngaphansi kwenqubomgomo kazwelonke yokuthuthukisa umnotho wedijithali ngamandla, i-cloud computing yase-China ingene enkathini yentuthuko ebandakanyayo, isidingo sekhompiyutha esisemaphethelweni sesikhulile, futhi ukuhlanganiswa konqenqema lwamafu kanye nesiphetho sekuyisiqondiso esibalulekile sokuziphendukela kwemvelo esikhathini esizayo.
Imakethe yekhompyutha yase-Edge izokhula ngo-36.1% CAGR eminyakeni emihlanu ezayo
Imboni yekhompyutha esemaphethelweni ingene esigabeni sokuthuthuka kancane kancane, njengoba kufakazelwa ukuhlukahluka kancane kancane kwabahlinzeki bayo besevisi, usayizi wemakethe okhulayo, kanye nokwandiswa okwengeziwe kwezindawo zokufaka izicelo. Mayelana nosayizi wemakethe, idatha evela embikweni wokulandela umkhondo we-IDC ikhombisa ukuthi usayizi wemakethe ophelele wamaseva wekhompyutha onqenqemeni e-China ufinyelele ku-US $ 3.31 billion ngo-2021, futhi usayizi wemakethe jikelele wamaseva e-edge computing eChina kulindeleke ukuthi ukhule ngokukhula okuhlanganisiwe konyaka. Izinga lika-22.2% kusukela ngo-2020 kuya ku-2025. U-Sullivan ubikezela usayizi wemakethe we-edge computing e-China kulindeleke ukuthi ufinyelele ku-RMB 250.9 billion ngo-2027, ne-CAGR engu-36.1% kusukela ngo-2023 kuya ku-2027.
I-Edge computing eco-industry iyachuma
I-Edge computing njengamanje isesigabeni sokuqala sokuqubuka, futhi imingcele yebhizinisi ochungechungeni lwezimboni ayilula. Kubathengisi ngabanye, kuyadingeka ukucabangela ukuhlanganiswa nezimo zebhizinisi, futhi kuyadingeka ukuba nekhono lokuzivumelanisa nezimo zebhizinisi kusukela ezingeni lobuchwepheshe, futhi kuyadingeka ukuqinisekisa ukuthi kunezinga eliphezulu ukuhambisana nemishini yehadiwe, kanye nekhono lobunjiniyela bomhlaba amaphrojekthi.
Uchungechunge lwemboni yekhompiyutha enqenqemeni luhlukaniswe ngabathengisi be-chip, abathengisi be-algorithm, abakhiqizi bemishini yehadiwe, nabahlinzeki bezixazululo. Abathengisi be-chip ngokuvamile bathuthukisa ama-chips e-arithmetic ukusuka ekugcineni kuya konqenqemeni kuya ohlangothini lwamafu, futhi ngaphezu kwama-chips aseceleni, baphinde bathuthukise amakhadi okusheshisa nezinkundla zokuthuthukisa isofthiwe. Abathengisi be-algorithm bathatha ama-algorithms okubona ngekhompyutha njengomongo wokwakha ama-algorithms ajwayelekile noma enziwe ngokwezifiso, futhi kukhona namabhizinisi akha izitolo ezinkulu ze-algorithm noma izinkundla zokuqeqesha neziphusha. Abathengisi bezinto zokusebenza batshala imali ngentshiseko emikhiqizweni yekhompiyutha enqenqemeni, futhi uhlobo lwemikhiqizo yekhompiyutha enqenqemeni luthuthukiswa njalo, kancane kancane lwakha isitaki esigcwele semikhiqizo yekhompyutha enqenqemeni ukusuka ku-chip kuya kuwo wonke umshini. Abahlinzeki bezixazululo bahlinzeka ngesoftware noma isoftware-hardware-integrated solutions ezimbonini ezithile.
Izinhlelo zokusebenza zemboni yekhompuyutha ziyasheshisa
Emkhakheni we-smart city
Ukuhlolwa okuphelele kwezakhiwo zasemadolobheni okwamanje kuvame ukusetshenziswa kumodi yokuhlola okwenziwa ngesandla, futhi imodi yokuhlola mathupha inezinkinga zezindleko ezithatha isikhathi esiningi nezidinga abasebenzi abaningi, ukuncika kwenqubo yomuntu ngamunye, ukumbozwa okungekuhle kanye nokuvama kokuhlolwa, kanye nezinga eliphansi. ukulawula. Ngesikhathi esifanayo inqubo yokuhlola irekhode inani elikhulu ledatha, kodwa lezi zinsiza zedatha azizange ziguqulwe zibe izimpahla zedatha zokunika amandla ibhizinisi. Ngokusebenzisa ubuchwepheshe be-AI ezimeni zokuhlola amaselula, inkampani idale imoto yokuhlola ehlakaniphile ye-AI yokuphatha emadolobheni, esebenzisa ubuchwepheshe obufana ne-inthanethi Yezinto, i-cloud computing, AI algorithms, futhi ephethe izinto zokusebenza ezichwepheshile njengamakhamera anencazelo ephezulu, ku- imibukiso yebhodi, namaseva aseceleni e-AI, futhi ihlanganisa indlela yokuhlola "yesistimu ehlakaniphile + umshini ohlakaniphile + usizo lwabasebenzi". Ikhuthaza ukuguqulwa kokubusa kwamadolobha kusukela ekugxileni kwezisebenzi kuye kobuhlakani bemishini, kusukela ekuhluleleni okunamandla ukuya ekuhlaziyweni kwedatha, kanye nokusuka ekuphenduleni kokungenzi lutho kuye ekutholakaleni okusebenzayo.
Emkhakheni wendawo yokwakha ehlakaniphile
Izixazululo zesiza sokwakha ezihlakaniphile ezisekelwe ku-Edge zisebenzisa ukuhlanganiswa okujulile kobuchwepheshe be-AI emsebenzini wokuqapha ukuphepha kwemboni yezokwakha, ngokubeka indawo yokugcina yokuhlaziya i-AI endaweni yokwakha, ukuqedela ucwaningo oluzimele nokuthuthukiswa kwama-algorithms we-AI abonakalayo asekelwe kuvidiyo ehlakaniphile. ubuchwepheshe be-analytics, ukutholwa kwezenzakalo zasikhathi sonke okufanele zibonwe (isb, ukuthola ukuthi kufanele yini ukugqoka isigqoko noma cha), ukuhlinzeka ngezisebenzi, indawo ezungezile, ukuphepha nezinye izindawo zokuhlonza izindawo ezinobungozi kanye nezinsizakalo zokukhumbuza i-alamu, kanye nokuthatha isinyathelo sokuqala Sokuhlonza abangaphephile. izici, ukuqapha okukhaliphile kwe-AI, ukonga izindleko zabasebenzi, ukuhlangabezana nezidingo zabasebenzi kanye nokuphepha kwezakhiwo zezindawo zokwakha.
Emkhakheni wezokuthutha ezihlakaniphile
I-Cloud-side-end Architecture isiphenduke ipharadigm eyisisekelo yokusatshalaliswa kwezicelo embonini yezokuthutha ehlakaniphile, nohlangothi lwamafu olunomthwalo wemfanelo wokuphatha okumaphakathi kanye nengxenye yokucubungula idatha, uhlangothi olusemaphethelweni luhlinzeka ngokuyinhloko ngokuhlaziywa kwedatha eseceleni kanye nesinqumo sokubala. -ukwenza ukucutshungulwa, kanye nohlangothi lokugcina ikakhulukazi olunesibopho sokuqoqwa kwedatha yebhizinisi.
Ezimeni ezithile ezifana nokuxhumana komgwaqo wemoto, ukuphambana kwe-holographic, ukushayela okuzenzakalelayo, nethrafikhi yesitimela, kunenombolo enkulu yamadivayisi ahlukahlukene afinyelelwe, futhi lawa madivayisi adinga ukuphathwa kokufinyelela, ukuphathwa kokuphuma, ukucutshungulwa kwe-alamu, nokusebenza nokulungiswa kokucubungula. I-Edge computing ingahlukanisa futhi inqobe, iguqule enkulu ibe encane, inikeze imisebenzi yokuguqulwa kwephrothokholi ye-cross-layer, ifinyelele ukufinyelela okuhlangene nokuzinzile, kanye nokulawula ngokubambisana kwedatha ehlukahlukene.
Emkhakheni wokukhiqiza izimboni
Isimo Sokuthuthukisa Inqubo Yokukhiqiza: Okwamanje, inani elikhulu lamasistimu okukhiqiza ahlukahlukene anqunyelwe ukungapheleli kwedatha, kanye nokusebenza kahle kwezinto zokusebenza nokunye ukubala kwedatha yenkomba kuwubudlabha uma kuqhathaniswa, okwenza kube nzima ukusetshenziselwa ukwenza kahle. Inkundla yekhompyutha ye-Edge esekelwe kumodeli yolwazi lwemishini ukuze kuzuzwe uhlelo lokukhiqiza lwezinga le-semantic lokuxhumana okuvundlile nokuxhumana mpo, okusekelwe endleleni yokucubungula ukugeleza kwedatha yesikhathi sangempela ukuze kuhlanganiswe futhi kuhlaziywe inombolo enkulu yedatha yesikhathi sangempela yasensimini, ukuze kuzuzwe umugqa wokukhiqiza osuselwe kumodeli. ukuhlanganiswa kolwazi lwemithombo yedatha eminingi, ukuze kunikeze ukwesekwa kwedatha okunamandla kokuthathwa kwezinqumo ohlelweni lokukhiqiza oluhlukile.
I-Equipment Predictive Maintenance Scenario: Ukugcinwa kwezinto ezisetshenziswa ezimbonini kuhlukaniswe ngezinhlobo ezintathu: ukunakekelwa kokulungisa, ukugcinwa kokuvimbela, kanye nokugcinwa kwezinto kusengaphambili. Ukugcinwa kokubuyisela esimweni sokubuyisela esimweni sangaphambili, ukugcinwa kokuvimbela, nokugcinwa kokubikezela kungokwesondlo sangaphambili, okokuqala kusekelwe esikhathini, ukusebenza kwemishini, izimo zesayithi, nezinye izici zokunakekelwa okuvamile kwemishini, kakhulu noma ngaphansi ngokusekelwe kumuntu. isipiliyoni, lokhu kwakamuva ngokuqoqwa kwedatha yezinzwa, ukuqapha kwesikhathi sangempela sesimo sokusebenza semishini, ngokusekelwe kumodeli yezimboni yokuhlaziywa kwedatha, nokubikezela ngokunembile ukuthi ukwehluleka kwenzeka nini.
Isimo sokuhlolwa kwekhwalithi yezimboni: inkambu yokuhlola umbono wezimboni iyifomu lokuqala lendabuko lokuhlola okuzenzakalelayo (AOI) endaweni yokuhlola ikhwalithi, kodwa ukuthuthukiswa kwe-AOI kuze kube manje, ekutholeni amaphutha amaningi nezinye izimo eziyinkimbinkimbi, ngenxa yokukhubazeka kwezinhlobonhlobo. kwezinhlobo, ukukhishwa kwesici akuphelele, ama-algorithms aguquguqukayo anwebeki kahle, umugqa wokukhiqiza ubuyekezwa njalo, ukufuduka kwe-algorithm akuguquki, nezinye izici, uhlelo lwe-AOI lwendabuko belunzima ukuhlangabezana nokuthuthukiswa kwezidingo zomugqa wokukhiqiza. Ngakho-ke, inkundla ye-algorithm yokuhlola ikhwalithi yezimboni ye-AI emelwe ukufunda okujulile + ukufunda isampula encane kancane kancane ingena esikhundleni sohlelo lokuhlola okubukwayo oluvamile, futhi inkundla yokuhlola ikhwalithi yezimboni ye-AI idlule ezigabeni ezimbili zama-algorithms wokufunda komshini wakudala kanye nama-algorithms okuhlola ukufunda okujulile.
Isikhathi sokuthumela: Oct-08-2023