Results - Stage 1

• A multi-tier architecture of an IoT platform, called EMULSION, has been elaborated as a representative of the new generation of IoT horizontal-type platforms.
• Initial implementation of the platform has been realized by integrating various (low-cost) electronic modules and (open-source) software components following the essential engineering principles for achieving the required interoperability. The main aim with EMULSION is to ensure the smooth operation of IoT devices of different types, accessible through heterogeneous (wireless) communication networks and cloud infrastructures, in order to provide a variety of personalized and contextualized services (by taking into account the current network context, service context, and the user preferences, by using big-data processing techniques) accessible via any mobile device (smartphone, tablet, laptop, etc.) anytime-anywhere-anyhow in accordance with the communication paradigm for always best connected and best-served users (ABC&S).
• A new vision has been elaborated for highly personalized and contextualized recommendations of services (delivered in real time) to mobile users, taking into account the current context. An entire system architecture has been proposed (for integration into EMULSION), which generates and dynamically handles user profiles for facilitating and optimizing the process of discovering and recommending services in order to achieve the best quality of service (QoE) for them. Based on this architecture, a prototype service will be developed, during stage 2 of the project, for the IoT areas of ‘smart (control of) environmental protection’ and ‘smart health’, which will be provided to users in the form of mobile app to find and recommend the "healthiest" route for their movement through areas with polluted air, i.e., routes with least risk to the health of the corresponding user-patient. The idea of the "healthiest" route was elaborated and proposed by the Principal Investigator of the project, Prof. Ivan Ganchev, and subsequently disseminated by him in the form of various articles and talks (including a plenary talk, two key talks, and an invited talk) at international scientific conferences and one national conference with international participation.
• A comprehensive complex model for intelligent recommendation of services to mobile users has been developed, based on a three-phase hybridization approach:
  (1) pre-filtering by using contextual information as to determine the initial set of relevant services for each particular user from the set of services available at the current time and satisfying the current context;
  (2) semantic recommendation by using mixed and weighted hybridization, and by taking into account additional factors, e.g., for matching and relevance;
  (3) post-filtering, based on additional information obtained about services and users.
• New hybrid models have been developed for the analysis and prediction (depending on weather conditions and other factors) of the average daily levels of fine particulate matter (PM2.5/PM10) in the air in order to prevent (and report in case of) exceeding the allowable thresholds of this major air pollutants for Bulgarian cities.