The aim of this master thesis is to show an electronic system infrastructure that isfully automated, whose final goal is to assess sleep disorders.Starting from the sensing elements and their firmware configuration based on the ap-plication, configuring parameters such as output data rate and full scale range, thisinfrastructure is able to acquire data and store them, using low cost devices which areless invasive then the standard used to perform this assessment.Substituting the standard with these devices two main benefits come from: the inva-siveness of the sensing elements during sleeping is reduced, and the data acquisition ismanaged by the hospital resources, so this infrastructure is scalable.Raw data will be transmitted through a standard protocol to a server station, in whichthe data analysis is performed.Statistical analysis and AI methods are used to analyze data.The first step is the data preprocessing, which is really important, especially whenheterogeneous data coming from different commercial devices have to be integrated.The goal of the analysis is to discover the different sleep stages, and this is done usingdifferent statistical methods that extracts the main features required to support thediagnosis.Also clustering methods are used in order to discriminate between sleep stages and,when possible, also the classification can be done, though it requires a labeled dataset.However, if the latter is available, classification models can be trained and may yieldbetter performance than clustering methods.Regarding scalability, the server station that makes the analysis has to be configured insuch a way that the streaming of data is allowed.