Vane ruzivo vevashandi vekuchera vanojairana nehunyengeri hunobvumira zviri mukati mebucket kutumirwa kupfuura kusvika kwebhoom. Robhoti yekuchera yakapihwa neural network yakave mudzidzi anoshingaira uyo zvakare kunanga kukanda matombo kusingasviki museve. Zvinotevera kukanda zvinhu zvakawanda uye kuvandudza iko kurongeka kwebasa remabhaketi pahurefu hwakasiyana.
Maitiro ekudzidzisa robhoti yekuchera neural network kuti ishandise nemazvo zviri mukati mebhaketi yakashumwa nevaongorori vanobva kuSwitzerland (ETH ZΓΌrich). Neural network yakavakirwa pakudzidza kwekusimbisa yakadzidziswa kukanda bhora nematombo pane imwe nzvimbo yakatarwa, iyo yaive kure nekusvika kwemuseve (kusvika 9,5 m nemuseve wekutora museve we7,5 m). Mashandiro akadaro achabatsira marobhoti kubata nehupamhi hwemabasa ane simba shoma rinoshandiswa pakufamba, uye achaitawo kuti basa rayo rive rakachengeteka.
Mucheri akaita kubata uye kukanda nebhaketi rine madhigirii maviri erusununguko, iyo yakanga isina kusimba yakagadziriswa kune boom. Kukanda zvakaitwa zvose mumutsara wakatwasuka, apo paingova nemuseve mumwe chete mukushanda, uye nekabhini inotendeuka. Muchiitiko chechipiri, kurongeka kwaive kushoma zvishoma, asi chero zvakadaro, projectile yakatsauka kubva painongedza nzvimbo isingapfuuri 30-40 cm.
Vatsvagiri vakadzidzisa neural network zvichibva pane yazvino 12-vhiri Menzi Muck M545 yekuchera. Kare, vakadzidzisa mucheki mumabasa akawanda asiri madiki, semuenzaniso, achidzidzisa kuvaka rusvingo rwakagadzikana kubva pamatombo asina kugadzirira. Mucheri pachake akaongorora kuenzana kwematombo uye akavaka fence yakasimba yematombo. Nekuita basa rakarurama rekuzvimiririra pasi, mucheri, achishandisa masensa akaiswa pairi, anovaka modhi yenzvimbo yakapoteredza iyo inoita mabasa akatarwa.
Source:
Source: 3dnews.ru
