Umdibaniso we ukudibanisa inzwa I-Software-defined vehicles (SDVs) iguqula indlela iimoto eziyilwe ngayo, zihlaziywa kwaye ziqhutywa njani. Sithetha ngendaleko apho i imithetho softwareI-orchestrates ezininzi zoluvo kunye ne-actuator, kwaye ivumela ukuphuculwa komsebenzi ngaphandle kokuchukumisa i-hardware.
Le ndlela, eyandayo ixhaphake kwicandelo, idibanisa iikhamera, i-radars, i-LiDAR kunye abenzi boluvo abanetial ukuqonda imo engqongileyo ngexesha lokwenyani, ngelixa i-architecture ephakathi kunye neqhagamsheleneyo yenza Uhlaziyo lwe-OTA, iinkonzo ezikude, imifuziselo emitsha yoshishino kunye neenkqubo zokuncedisa umqhubi eziphucukileyo (ADAS) kunye nemisebenzi yokuqhuba ezenzekelayo.
Yintoni isithuthi esichazwe ngeSoftware?
I-VDS sisithuthi esisebenza kakuhle ichazwe kwaye ilawulwa yisoftwareNgokungafaniyo neemoto zemveli, apho inkqubo nganye ye-elektroniki kunye neyomatshini isebenza ngokungqongqo kwaye yodwa, ii-SDVs zilungelelanisa yonke imisebenzi ngamaqonga anamandla ekhompyuter, unxibelelwano lwangaphakathi nangaphandle, kunye noluhlu lwesicelo oluguqukayo ngokuhamba kwexesha.
Le ngcamango ayizange ivele ekuqaleni; amandla ayo avela ekuxhumeni ngaphakathi abenzi boluvo, ii-activators, ii-hardware ezisebenza kakhulu kunye ne-algorithms (kubandakanya i-AI kunye nobunjineli bokulawula) esele ivumela ubudlelwane obunzulu phakathi kwe-traction, braking, steering, perception, navigation kunye namava omsebenzisi.
Umahluko xa kuthelekiswa nezithuthi zemveli
Kwimoto eqhelekileyo, imisebenzi emininzi ixhomekeke kuyo hardware ethile nokungaguquguquki. Kwi-VDS, undoqo yisoftware, eyandisa unxulumano phakathi kwe-subsystems kwaye yenza lula ukuvela kwemisebenzi kumjikelezo wobomi besithuthi.
Uphuculo alusafuni kutshintshwa iindawo: zihanjiswa nge uhlaziyo olukude Ezi zikhawuleza, zinciphisa iindleko kunye nexesha, nangona zibandakanya ukhuseleko olusebenzayo kunye nemingeni ye-cybersecurity. Esi sikhundla sohlaziyo oluqhubekayo luvumela ukubandakanywa kobuchwepheshe bokusika kunye izincedisi zokhuselo ngobuchule.
Ukongeza, ii-SDVs zinako qokelela idatha ngexesha lokwenyani ukwandisa ukusebenza, ukukhulisa iinguqulelo zexesha elizayo, kunye nokukhawulezisa ukuveliswa kwezinto ezintsha kunye neenkonzo.
SDV Architecture
Isiseko somzimba siquka Inkqubo yokusebenza okuphezuluuthungelwano lwangaphakathi, ukugcinwa, izinzwa ezininzi, kunye ne-activators ezisasazwayo. Eli qonga le hardware eyomeleleyo Isebenzisa isoftware yexesha lokwenyani, idibanisa amacandelo, kwaye ingqamanisa imiyalelo kwiziqhoboshi, ukuqhubela phambili, okanye ukuqhubela phambili.
Ngokusekwe kuloo nto, isoftware icwangciswe ngokwemigangatho: a inkqubo yokusebenza ilawula i-cybersecurity, imemori, kunye ne-I / O; umaleko we i-virtualization okanye i-middleware Idibanisa kwaye ibeka umgangatho wonxibelelwano; kunye nomaleko we sicelo Isebenzisa imisebenzi (i-traction, i-braking, i-steering, njl.) ngaphandle kokuxhomekeka kwiinkcukacha ze-hardware ephantsi.
Uyilo lwezakhiwo lugqityiwe ngoqhagamshelwano lwangaphandle: i-Intanethi, 5G, V2X kunye nelifu leenkonzo ezikude, ukuxilongwa, uhlalutyo kunye nokuthunyelwa kohlaziyo lwe-OTA kwizithuthi kunye neenqwelo.
Amaqonga ophuhliso kunye ne-ecosystems
Ukubekwa emgangathweni kunye nentsebenziswano ngundoqo. I-AUTOSAR, kwiiprofayili zayo Classic (ixesha lokwenyani, ukhuseleko kunye nokuthembeka okuphezulu) kunye I-Adaptive (iindawo eziguquguqukayo, uhlaziyo lwe-OTA kunye noqhagamshelo), ibe sisiseko sokusebenzisana kwishishini leemoto.
Izixhobo kunye neenkqubo ezineendlela ezikhethekileyo nazo ziyazuza: i Rust ulwimi kukhuseleko kunye nokusebenza kakuhle kwememori; kunye ne-RTOS njenge IiFreeRTOS (umthombo ovulekileyo, izixhobo ezinobuncwane) kunye SAFERTOS (isiqinisekiso sokhuseleko olusebenzayo), ilungele amacandelo abalulekileyo.
Abathengisi beSilicon kunye nesoftware banikezela ngeemeko ezizinzileyo zee-SDVs ezineprosesa ezikhawulezayo eziqhubayo. ukusetyenziswa kwakhona kwekhowudiIzilawuli zomgangatho wemveliso kunye ne-MCAL ejolise kukhuseleko. Bakwanikezela ngeenketho ze-OS ezifana I-FreeRTOS, i-Linux, i-QNX kunye ne-SafeRTOSkunye nokuhambelana ne-AUTOSAR ukwenza lula ukudityaniswa ne-ecosystem eyahlukeneyo.
Imbali kunye nokuzivelela kwezinto
Ukususela ngo-70, iinkqubo zokuqala zombane zilawulwa injini kunye nezinto ezikhutshwayoKwiminyaka yoo-90s, ii-microprocessors zenza ukuba kube lula ukulawula ukutsalwa kunye ne-ABS, kwaye ngewaka leminyaka elitsha kwafika unxibelelwano, idijithali, kunye... I-ADAS, ukwandisa ukhuseleko kunye nentuthuzelo.
Namhlanje, uncedo oluphambili kunye nezixhobo ezizenzekelayo zidibanisa izinzwa, iikhamera, kunye ne-algorithms yexesha lokwenyani, ezisa amanqanaba ahlukeneyo ukuqhuba ngokuzimela ngombono obanzi wesithuthi njengenkqubo.
Ukudityaniswa kwenzwa kwii-SDVs
Udibaniso lwesivamvo ludibanisa idatha esuka kwimithombo emininzi ukunika imbono ebanzi ichanekile, yomelele kwaye iluncedo yokusingqongileyo kunye nesithuthi ngokwaso. Ivumela ukoyisa imida yenzwa yodwa ngokudibanisa imiqondiso ebonwayo, umgama kunye nentshukumo kunye neemodeli zemathematika kunye ne-AI.
Amacandelo ayo abalulekileyo abandakanya: i ukubamba usebenzisa iikhamera, i-LiDAR, i-radar, i-ultrasonic kunye ne-inertial sensors; i iqhutyelwe phambili (ukucoca, ungqamaniso, kunye nokulungelelanisa); i ukudibanisa algorithms ezidibanisa imiqondiso; kunye nenqanaba le ukwenza izigqibo okondla ulawulo kunye nocwangciso.
I sebenza njani
Okokuqala, i-heterogeneous data iqokelelwa, emva koko ilungelelaniswe kwaye ihluzwe ukunciphisa ingxolo kunye ne-bias, kwaye ekugqibeleni idibaniswe nobuchule obubuyisela imeko ehambelanayo yokusingqongileyo. Lo mboniso utolikwa ukuze uqalise imisebenzi efana ne ukufumanisa umqobo, ukulandelwa kwezinto okanye iintshukumo.
Phakathi kwezona zindlela zixhaphakileyo zi Isihluzo seKalman Ukuqikelela amazwe anengxolo, iindlela zeBayesian zokuhlaziya izinto ezinokwenzeka ngobungqina obutsha kunye Ukudibanisa okusekwe kufundo olunzuluapho iinethiwekhi ze-neural zifunda ukudibanisa iisignali ezininzi.
Imingeni yobugcisa
La ungqamaniso lwexeshana Phakathi kwabenzi boluvo abanezakhelo ezahlukeneyo kunye nezitampu zexesha, ulungelelwaniso olomeleleyo kunye nezicwangciso zokutywinwa zethutyana ziyafuneka.
El ingxolo nokungaqiniseki Azinakuphepheka: izihluzi, imifuziselo enokwenzeka, kunye nokulinganisa rhoqo ziyafuneka ukugcina umgangatho wedatha.
La ukuntsokotha kokubala Iphezulu, ngakumbi ngexesha langempela; edge computing kunye hardware acceleration uncedo ukuqulatha latencies.
Yila i ukuhambelana Ukuphepha ukuphindaphinda okungqubuzanayo kunye nokusombulula ukungangqinelani phakathi kweenzwa ngumngeni obalulekileyo wokwakha.
Izicelo
Kwizithuthi ezizimeleyo kunye ne-ADAS, i-fusion ixhasa i ukuhambaImbono ye-360 ° kunye nocwangciso lwendlela. Kwirobhothi, iququzelela ukukhohlisa kunye nendawo; kwizixeko ezihlakaniphile, idibanisa izibonakaliso ze-IoT zokuhamba kunye namandla; kukhathalelo lwempilo, izixhobo ezinxitywayo zidibanisa iimetriki ezininzi; nakwishishini, iyaqhuba Ulondolozo lwangaphambili kunye nolawulo lomgangatho.
Iinzuzo kunye noqhagamshelwano
Ukhuseleko luqiniswa yi I-ADASEzi nkqubo zinciphisa imingcipheko ngokubona okuphambili kunye nolawulo olusasazwayo. Benza iimpendulo ezilungelelanisiweyo kwi-braking, ukuqondisa, kunye nokukhawulezisa ngamaxesha okuphendula angenakwenzeka kumntu.
- i-adaptive cruise control: Lungisa isantya ukugcina umgama.
- Uncedo lokupaka: uncedo kubuchule bokwenza izinto zoluvo kunye neekhamera.
- Ukunyanzeliswa kwesiqhoboshi sikaxakeka: thatha isenzo xa kukho ingozi yongquzulwano.
- Ukugcinwa kwendlela/utshintsho: inqanda ukutenxa kwaye ixhasa i-maneuver.
- Ukubona indawo eyimfama: isilumkiso malunga neendawo ezifihliweyo.
Ngokumalunga nokusebenza kakuhle, ii-SDVs zivumela ulungiselelo oluqhubekayo Ngokusekwe kwidatha yemoto kunye nokusingqongileyo, ngokubeka iliso okude, ukuxilongwa kwangaphambili kunye nokumisa iworkshop ezimbalwa.
Kubuntu, abasebenzisi bavula izinto ezifunwayo kwaye bafumane uphuculo. I-OTANangona kunjalo, yayiyilwe ngononophelo ukuthobela izithintelo zokhuseleko kunye nokuphepha umngcipheko ngexesha lohlaziyo.
Uqhagamshelo olukwinqanaba eliphambili lenza iinkonzo ezifana nokuhamba ngexesha lokwenyani, ulawulo lwezithuthi, ukuzonwabisa, kunye V2Xukuguqula amava asebhodini kunye nobudlelwane phakathi kwesithuthi, iziseko zophuhliso kunye nelifu.
Isishwankathelo semarike kunye neemodeli zoshishino
Ukutshintshela kwi-centralized computing kunye ne-quasi-zonal architectures kuqhuba ixabiso. Ezi qonga ziqikelelwa ukuba zivelise ngeenxa zonke 755.000 iidola ezigidi kwingeniso hardware ngo-2029, ngelixa imisebenzi ye-SDV iya kukhula ngesantya se 30–34% ngonyaka kude kube ngu-2035 enkosi ekwenzeni imali ngeenkonzo ezidibeneyo nezizimeleyo.
Ii-SDVs zinokuhlelwa zibe ngamanqanaba amahlanu, ukusuka kuyilo olugxile kwiindlela zokuhamba kunye nemimandla ukuya kwizithuthi ezidityanisiweyo ngokupheleleyo. software-centricEmbindini kukho i-HPC, abalawuli bendawo kunye ne-middleware ejolise kwinkonzo, eyenza ukwahlulwa kwe-hardware-software kunye nokulinganisa okusebenzayo.
Ishishini lihlengahlengiswa nge iimpawu ezifana nenkonzo, urhwebo lwemoto kunye ne-cockpit yedijithali apho i-AI kwisixhobo (kunye nabadlali abafana ne-Qualcomm, i-Nvidia okanye i-Unity) yenza amava aguquguqukayo: izikrini ezigcweleyo, ii-avatars ze-AI kunye "nezikhumba" ezinokwenziwa ngokwezifiso.
La Uqhagamshelo lweV2X (C-V2X, DSRC, kunye ne-5G) isitshixo sokhuseleko kunye nokulungelelanisa; ukwamkelwa kwayo kuxhomekeke kwi-spectrum kunye nemigaqo-nkqubo ngommandla (eChina, EU, US, Japan, Korea). Ukudityaniswa kwee-OBUs, ii-RSU, kunye ne-chipsets zihambelana namaqonga e-SDV ukukhawulezisa ukuthunyelwa.
Imingeni engaphaya kobugcisa
Iimodeli zentlawulo ezisekelwe kumrhumo zinokuvelisa ukwaliwa kwabathengi ukuba zisetyenziswa kwimisebenzi ebonwa njengomgangatho, echaphazela umbono wophawu.
Uqhagamshelwano olukhulu luzisa imingcipheko yokhuseleko lwe-intanethi Ngokumalunga nokulawulwa kwesithuthi, ubumfihlo, kunye neenkonzo zefu, izikhokelo eziphambili kunye nokubeka iliso okuqhubekayo kuyafuneka.
La ubunini kunye nokhuseleko lwedatha Ifuna imigaqo-nkqubo ecacileyo yokugcina, ukusetyenziswa kunye nokwabelana, kunye nemvume kunye nokuthotyelwa kokulawula.
Uphuhliso, ukuqinisekiswa, kunye nokugcinwa kwamaqonga e-SDV kubandakanya iindleko eziphezulungokukodwa kwimisebenzi ebalulekileyo kunye neziseko zokuhlaziywa kwe-OTA ekhuselekileyo.
Ubunzima butshintshela kwizigidi zemigca yekhowudi, iileya ezininzi, kunye nabathengisi, ukwandisa umngcipheko iimpazamo zokuhlanganisa kunye nokusilela.
Zikhona ukunqongophala kwetalente kwisoftware, i-AI, kunye ne-cybersecurity ngaphakathi kwabavelisi abanenkcubeko eyomeleleyo yoomatshini; abaninzi abalindelanga ukugqiba izakhono zangaphakathi ngaphambi kweminyaka elishumi ezayo.
Vuka imingeni yolawulo Ngokuphathelele uxanduva, uhlaziyo lokuziphatha kwesoftware, kunye nokhuseleko oluvelayo, ngakumbi ngemisebenzi ezenzekelayo.
I-AI iphakamisa imibuzo malunga ukucaciswa kunye nokuqikelelwakunye nolawulo lokurhoxiswa okanye iimeko zokuphela konyaka ekuqhubeni ngokuzenzekela.
La ukwahlulwa Iiplatform, iinkqubo zokusebenza, kunye namafu anzima ukuhambelana kunye nokulinganisa phakathi kweemodeli kunye nemimandla.
Nangona ii-OTA zifanelekile, enye ulawulo lohlaziyo olubi Oku kunokubangela ukusilela kwenkqubo kunye nokuphazamiseka komsebenzisi; ulawulo novavanyo lubalulekile.
Iimanyano kunye neendlela zoshishino
I-Bosch kunye neCariad yomeleza intsebenziswano yabo kwiNqanaba lesi-2 kunye ne-3 yokuncedisa kunye nemisebenzi yokuqhuba ngokuzenzekelayo ngephakheji yesofthiwe esekelwe IABaphuhlisa onke amacandelo ngokuzimeleyo, befuna ukuziphatha njengendalo njengaleyo yomqhubi womntu kunye nokhuseleko oluphezulu.
Iimpawu zokuqala sele zivavanywa kwiinqwelo zokulingwa kwaye ziqeqeshwe ngemithamo emikhulu yedatha. Injongo kukuba nephakheji efanelekileyo imveliso ukusuka phakathi ku-2026, edityanisiweyo kuyilo olutsha lwe-SDV yeQela leVolkswagen kwaye inobungozi kwabanye abavelisi.
I-AI isetyenziswa kulo lonke ikhonkco: ukuqonda, ukudityaniswa kweekhamera kunye neeradaUkwenza izigqibo kunye nolawulo olukhuselekileyo lwe-powertrain, i-steering, kunye neziqhoboshi. Ukujonga phambili, iindlela zemultimodal ziyaphononongwa. Umbono-uLwimi-Isenzo ukuqiqa malunga neemeko ezinzima kwaye ufumanise umngcipheko ofihlakeleyo.
Ulawulo olupheleleyo lwekhowudi yomthombo kunye nepropathi enomgangatho ophezulu wokuqonda ivumela ukumiselwa kwemigangatho ephezulu ukhuseleko lwedatha, ukhuseleko kunye nokungafihlikunye nezigqibo ze-AI ezinokulandeleka kwaye zichaze. Ubunjineli buxhaswa sisicwangciso se-hardware esinokwehla kuzo zonke izintlu.
Uvavanyo lwenziwa kwiindlela zikawonke-wonke eYurophu, eJapan nase-USA, kunye nezithuthi ezifana ID.Buzz y Audi Q8Kulo nyaka, amakhulu eenqwelo ezixhotyiswe ngee-suite ze-sensor ezipheleleyo zongezwa ukuba zibambe iimeko ezinqamlekileyo; uphuhliso lu data-qhutywangophuculo lwemihla ngemihla.
Izibonelelo kunye nokufunda
Ngolwazi oluthe kratya kubalawuli abangundoqo abenza i-SDV, kuyanceda ukuphonononga uxwebhu lobugcisa lwe-Aptiv. Ukufikelela ngokuthe ngqo: Landa i PDF, apho kuchazwe indlela ulwakhiwo olusembindini oluphucula ngayo ukusebenza okuphezulu kwekhompyuter, izowuni kunye nohlaziyo oluqhubekayo.
Imisebenzi kunye nemithombo yereferensi kwi-sensor fusion kunye noqikelelo: i-Durrant-Whyte kunye ne-Bailey (SLAM), i-Thrun/Burgard/Fox (i-robotics enokwenzeka), u-Bar-Shalom et al. (ukulandelela kunye nokuhamba); ukongeza kwizibonelelo zemfundo evela NVIDIA malunga ne-sensor fusion yokuzimela kunye Intel edge computing isetyenziswe kwesi sihloko.
Ukujonga umfanekiso uphela, indibaniselwano ye-SDV kunye ne-sensor fusion yenza ukuba ungene ngaphakathi ukhuseleko, impumelelo kunye namavaIqonga eliphilayo elifunda kwiidatha, uhlaziyo ngaphandle kokutshintsha i-hardware, kwaye livula iimodeli zoshishino ezixhunyiwe, ngaphandle kokuba i-cybersecurity, umgangatho wesoftware, kunye nokuthembela kwi-AI kulawulwa ngokungqongqo.
