Abstract
Forecasts announce that the number of connected objects will exceed 20 billion by 2025. Objects, such as sensors, drones or autonomous cars participate in pervasive applications of various domains ranging from smart cities, quality of life, transportation, energy, business or entertainment. These inter-connected devices provide storage, computing and activation capabilities currently under-exploited. To this end, we defined “Spatial services”, a new generation of services seamlessly supporting users in their everyday life by providing information or specific actions. Spatial services leverage IoT, exploit devices capabilities (sensing, acting), the data they locally store at different time and geographic locations, and arise from the spontaneous interactions among those devices. Thanks to a learning-based coordination model, and without any pre-designed composition, reliable and pertinent spatial services dynamically and fully automatically arise from the self-composition of available services provided by connected devices. In this paper, we show how we extended our learning-based coordination model with semantic matching, enhancing syntactic self-composition with semantic reasoning. The implementation of our coordination model results in a learning-based semantic middleware. We validated our approach on various experiments: deployments of the middleware in various settings; instantiation of a specific scenario and various other case studies; experiments with hundreds of synthetic services; and specific experiments for setting up key learning parameters. We also show how the learning-based coordination model using semantic matching favours service composition, by exploiting three ontological constructions (is-a, isComposedOf, and equivalentTo), de facto removing the syntactic barrier preventing pertinent compositions to arise. Spatial services arise from the interactions of various objects, provide complex and highly adaptive services to users in seamless way, and are pertinent in a variety of domains such as smart cities or emergency situations.
Highlights
In his seminal and visionary paper Mark Weiser [1] describes the notions of “disappearing computers and disappearing technology”, where connected objects are “embedded in the everyday world” and made invisibly available to the users
We propose to use a reinforcement learning module in each agent to learn the best action to take regarding the partial schema that it receives during the composition
Spreading: Our strategy to coordinate between all available nodes in the network is to spread a copy of the Live Semantic Annotations (LSA) every time a new property is added in the property list
Summary
In his seminal and visionary paper Mark Weiser [1] describes the notions of “disappearing computers and disappearing technology”, where connected objects are “embedded in the everyday world” and made invisibly available to the users. The latter “use them unconsciously to accomplish everyday task”, such as obtaining information, requesting actions, or animating otherwise inert objects-in other words, providing an “embodied virtuality”. Thirty years later, connected objects are dramatically increasing in number, and more than 20 billion more objects are expected to be connected in the five years They relate to many domains such as smart city, smart building, transportation, energy, business, or quality of life.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
