| Popis: |
V okviru diplomske naloge smo izdelali hierarhični sistem za avtomatizacijo, ki je sestavljen iz osrednjega strežnika ter posameznih povezanih odjemalcev. Delovanje obeh je vzajemno, saj se med njima vseskozi izmenjujejo zahteve, ki opisujejo stanja nanj priključenih zunanjih naprav. Z njimi lahko upravljamo preko uporabniškega vmesnika ali z glasovnim upravljanjem na strani odjemalca. Uporaba slednjega nam omogoča upravljanje naprav s pomočjo govora, kjer se nam kot povratni odziv predvaja odgovor v obliki umetno tvorjenega govora. Naš končni izdelek je sestavljen iz spletne in odjemalne aplikacije, z uporabo katerih lahko v eno skupno mrežo povežemo različne tipe zunanjih naprav ali senzorjev. Za primarni preizkus odjemalne aplikacije smo uporabili Raspberry Pi računalnik, ki nam vse to omogoča s svojimi vgrajenimi GPIO nožicami. Z napravami lahko upravljamo ročno preko vmesnika ali spletni aplikaciji določimo časovni termin, v katerem naj ta samodejno aktivira izbrane naprave. Pri razvoju glasovnega upravljanja smo za razpoznavo ter sintezo govora uporabili storitvi Google Speech Recognition ter Google Synthesis. Z razpoznanim govorom v tekstovni obliki lahko nato upravljamo s priključenimi napravami ali ga pošljemo storitvi Wolfram Alpha, ki poskuša zanj poiskati primeren povratni odziv. Ta se po opravljenem postopku sinteze predvaja na odjemalcu preko priključenih zvočnikov. In the thesis, we have developed a hierarchical system for automation, which consists of a central server and connected clients. The operation of the two is mutual because they constantly exchange requests with latest states of connected devices. They can be managed through the user interface or by voice control on the client side. With the recognized speech, we can control specific connected devices and, as a result, get a response in the form of synthesized speech. Our final product is composed of web and client applications that can connect different types of external devices or sensors in the unified network. For initial test of the client application, we decided to use the Raspberry Pi computer, which enables connection of external devices to its built-in GPIO pins. Those devices can be managed manually via the user interface or by web application – user can set the time when the application should automatically activate the selected devices. In the development of the voice control interface, internet services Google Speech Recognition and Synthesis were used. Using recognized speech content, we can manage connected devices or send text to service Wolfram Alpha it tries to find an appropriate answer, that can be played as a synthesized speech on client using connected speakers. |
