Nekuda kwemhando dzakasiyana dzezvidzikisira uye mota dzinoshandiswa mune dzakasiyana dzekushanda pamusoro pecheni dhizaini maconveyors, iyo inopindirana yekumisikidza sensor ichachinjawo. Naizvozvo, sarudza nzvimbo yekumisikidza yeiyo inoderedza sensor mushure mekunyatsoongorora. Nekuda kweiyo yakasarudzika nharaunda yekushanda pamusoro pecheni dhizaini yekutakura, iyo sensor ichave yakabonderana kana kukuvadzwa. Kuti uve nechokwadi chekuti sparks dzakagadzirwa kana sensor yakuvadzwa (kunyanya inoreva mutsara wechiratidzo che sensor uye wedunhu kuburitswa uye kubuda kunze), hazvizokonzerese iyo sensor kwairi. Kana kuputika kukaitika munzvimbo inoputika gasi, zvese sensor magetsi uye chiratidzo chekufambisa chinoda kuzadzisa zvinodiwa zvekuchengetedza. Kureva kuti, iyo sensor pachayo inofanirwa kunge iri intrinsically yakachengeteka sensor, uye magetsi eiyo sensor anofanirwa kusangana nemukati zvakachengeteka zvinodiwa.

Mhosva yekuongororwa ndeyekutonga mamiriro ekushanda kana mamiriro asina kunaka emutakuri weketani. Rine zvirevo zviviri. Imwe ndeyekufanotaura uye kufanotaura mamiriro ekushanda emidziyo yekufambisa isati yatangisa cheni; imwe yacho ndeyekufanotaura nezvenzvimbo, chikonzero, rudzi uye chiyero chekutadza mushure mekunge michina yakundikana. kutonga uye kuita zvisarudzo zvekuchengetedza. Mabasa ayo makuru anosanganisira kuona kukanganisa, kuzivikanwa, kuongorora, kufungidzira uye kuita sarudzo. Nzira dzekuongorora mhosva dzinosanganisira zvikamu zviviri: nzira dzekuongorora mhosva dzinobva pamhando dzemasvomhu uye nzira dzekuongorora mhosva dzinobva pahungwaru hwekugadzira. Iyo yekutadza kuongororwa nzira yakavakirwa neural network uye ruzivo fusion tekinoroji inotsanangura iwo musimboti weneural network uye ruzivo fusion. Panguva imwecheteyo, mienzaniso yekukanganisa kuongororwa kwakavakirwa neural network uye kukanganisa kunoenderana neumboo dzidziso inopihwa.

Iyo neural network yechain plate conveyor inogona kukamurwa kuita mapoka maviri zvichienderana nenzira dzakasiyana dzekubatanidza pakati peeuroni: mhinduro-yemahara yemberi network uye mutual musanganiswa network. Iyo mhinduro-yemahara yekumberi network ine yekuisa layer, yepakati layer uye yekubuda layer. Rutivi rwepakati runogona kuumbwa nezvikamu zvakati wandei, uye nyuroni dziri muchikamu chega chega dzinogona kugamuchira chete kubuda kwemaneuroni muchikamu chakapfuura. Panogona kuve nekubatana pakati pechero neuroni mbiri mune yakabatana network, uye chiratidzo chekupinza chinofanira kudzokororwa chichiendeswa mberi nekudzoka pakati peeuroni. Mushure mekuchinja kwakati wandei, iyo cheni inotakura inoenda kune imwe nzvimbo yakagadzikana kana kupinda periodic oscillation uye imwe nyika.