Networking and Managing Smart Campus Devices Using the Home Assistant Platform

U ovom je radu opisano rješenje pametnog kampusa koje omogućava pristup umreženim uređajima ovisno o pravima korisnika te detekciju broja ljudi u prostoriji za izvedbu automatiziranog upravljanja uređajima. Sustav definira studentima i profesorima pravo paljenja i gašenja pametne lampe, a isključivo profesorima omogućuje promjenu boje svjetla lampe. Autorizacija pristupa ostvarena je koristeći tokene, gdje članovi svake grupe imaju pristup samo tokenu te grupe. Detekcija broja ljudi je odrađena koristeći kameru s unaprijed istreniranom neuronskom mrežom koja očitava broj ljudi na slici. U slučaju preopterećenja kapaciteta broja ljudi u prostoriji, boja se svjetla promjeni u crveno. Sustav se temelji na uređajima umreženim pomoću platforme Home Assistant na koju je spojen i alat Node-RED pomoću kojeg su definirani svi tokovi u kojima se provodi funkcionalnost sustava, te Nvidia DeepStack alatu koji prepoznaje broj ljudi na slici.This thesis describes a smart campus solution that provides access to networked devices depending on the rights of users, as well as detection of the number of people in a room for the execution of an automation that controls a device. The system grants students and professors the right to turn a smart lamp on or off, and the right to change the color of the lamp’s light is granted exclusively to professors. Access authorization is achieved using tokens, where the members of each group have access only to that group’s token. The detection of the number of people in a room is done using a camera with a pre-trained neural network that counts the number of people in the picture. In case of overloading the room capacity of the number of people, the color of the light changes to red. The system is based on devices networked using the platform Home Assistant, through which the Node-RED tool is integrated, for defining all the flows through which the system functionality is carried out. The Nvidia DeepStack tool is used to detect the number of people in a picture