Search

V diplomskem delu smo obravnavali izdelavo pogovornega bota oz. asistenta za komunikacijo s sistemom pametne hiše. Na začetku dela smo opisali trenutno stanje avtomatizacije domov ter predstavili, kaj sploh je pogovorni bot, njihovo zgodovino in uporabo pogovornih botov danes. Nadaljevali smo s predstavitvijo spletne platforme IBM Cloud. To je platforma podjetja IBM, kjer podjetje svojim strankam ponuja oblačne storitve. Nekatere storitve so na voljo tudi v brezplačnih različicah in ena izmed teh je IBM Watson Assistant, ki smo ga uporabili za izdelavo pogovornega bota. To je storitev, ki uporabnikom omogoča izgradnjo in uporabo pogovornega bota in je podrobneje predstavljena v nadaljevanju dokumenta. Sledila je še predstavitev načrtovanja hišnega asistenta, kjer smo predstavili samo idejo in opisali namišljen objekt oz. hišo, za upravljanje katerega/katere smo izdelali pogovorni bot. V naslednjem delu smo opisali potek izdelave pogovornega bota, njegove glavne sestavne dele, način programiranja in izzive, s katerim smo se soočili pri programiranju. Na koncu smo predstavili še ugotovitve, in sicer na nivoju pravilnosti delovanja izdelanega asistenta, njegove robustnosti, dejanske uporabnosti pri uporabi za komunikacijo z zalednim sistemom pametne hiše in možnosti za nadaljnji razvoj rešitve. In this thesis, we discussed creation of chatbot assistant for communication with smart-house's backend system. At the beginning, we described home automatisation as-is situation, introduced chatbots in general, described their history and use of chatbots today. We continued with introduction of IBM Cloud platform. IBM Cloud is IBM's platform, where company is offering their clients their cloud-based services. Some of those services are also offered in Lite plans (free versions). One of those is Watson Assistant, which we have used to build our chatbot. Watson Assistant is cloud service that enables users to create and use chatbot. The assistant is further presented in detail later in the thesis. Following part of the thesis is presentation of our home assistant planning, where we presented the idea of home assistant and described how chatbot is positioned within complete home assistant solution. Later we also presented programming process, it's main components, programming method and challenges we faced during the programming. At the end we also described key findings, both at the level of chatbot's correctness of understanding user's inputs and actual usability for communication between home assistant and backend system. We also discussed ways for further improvement and development of our built chatbot.

V magistrskem delu je predstavljen sistem pametnega domovanja, krmiljen s pomočjo odprtokodne platforme Home Assistant. Sistem je razdeljen na medsebojno neodvisne podsisteme namenjene nadzoru in upravljanju, osvetlitvi, ogrevanju, multimediji in senčenju. Integrirano spremljanje in krmiljenje domovanja lahko poteka na daljavo s pomočjo osebnega računalnika, mobilnega telefona ali z glasovnimi ukazi v slovenskem jeziku. Predstavljena je izdelava prilagodljive osvetlitve Hyperion Ambilight ter njena integracija v sistem. Za dviganje in spuščanje senčil je zasnovan sistem s servomotorjem in mikrokrmilnikom. Vsi podsistemi so povezani s pomočjo platforme Home Assistant, ki se izvaja na mikroračunalniku Raspberry Pi 3. Pametno domovanje je uspešno realizirano in demonstrirano v stanovanjskem objektu. The master's thesis presents smart home system, controlled by an open-source platform Home Assistant. The system is divided into independent subsystems designed for monitoring and control, lightning, heating, multimedia and shading. Integrated monitoring and control of a smart home can be done remotely from a personal computer, a mobile phone or via voice control in Slovenian language. Hyperion Ambilight adaptive lightning and its integration into the system is presented. A system for controlling the blinds is designed based on an electric servomotor and a microcontroller. The independent subsystems are connected by the Home Assistant platform running on Raspberry Pi 3. Smart home is successfully implemented and demonstrated in a home apartment.