SmartSPEC is a smart space simulator and data generator that creates customizable smart space datasets using semantic models of spaces, people, events and sensors. We employ ML-based approaches to characterize and learn attributes of the embedded people and events in a sensorized space and apply an event-driven simulation strategy to generate realistic simulated data about the space (events, trajectories, sensor datasets, etc.).
The SmartSPEC architecture consists of two main components:
Scenario Learning, which uses input seed connectivity data and a priori knowledge of the underlying space and sensors to learn high-order concepts of events, people and trajectories, which we refer to as “metaevents”, “metapeople” and “metatrajectories”, respectively; and
Scenario Generation, which takes SmartSPEC data to generate a synthetic dataset from which a smart space dataset (e.g., trajectory dataset, sensor observation dataset, etc.) can be derived. We use the variations of the data models described below to define various scenarios, which drives the generation of new observable phenomena in the smart space.
Useful links and references
Roberto Yus
Assistant Professor
My research interests include Data Management, Knowledge Representation, the Internet of Things, and Privacy.
Publications
SmartSPEC: A framework to generate customizable, semantics-based smart space datasets
