Mist Computing is an architectural paradigm in which some of the operations of an IoT network are moved towards the nodes on its edge, in order to fully utilize their increase in computational power and to analyze the type of advantages under this structure. This work is a study aimed at implementing this type of paradigm in an indoor setting, in order to gather information about the activities related to the environment; this information can be used to give orders to different actuators nodes.
Several application protocols have been analyzed, assessing their positive and negative aspects and taking into account the mostly used standards in the community. By creating an environment in which there is a bond between video and audio data, smart devices are bound to everyday events. This allows to obtain information on everyday habits and collect them in statistics in order to build activity schemes for the actuators.
Some architectural schemes have been proposed for the modeling and implementation of the chosen paradigm nodes, putting them in the general contest of Internet of Things and comparing them with other common paradigms used. The study also focuses on a particular kind of node, that acquires and analyzes audio data; these pieces of information obtained through the node play a pivotal role in the project and mark the starting point for the analysis of the context taken into consideration. The strength of this approach lies on the possibility of using this type of application with remote data analysis services, capable of constructing recurrent behavioral schemes to be matched to related activity schemes for the actuators.