Covering our world with sensors

Abstract

Information technology has had profound impacts on our lives. The problem is that, so far, technology has required our explicit attention to provide services. This limits the scenarios in which it can or we would like it to take action. On the other hand, perceptive systems aim to understand our activities and intentions to proactively, collaboratively, and adaptively provide services. This requires systems to form projections of the world, but also construct models for how to respond. This talk starts by looking at how deploying large numbers of form-fitting sensors, which are explicitly associated with the physical objects we interact with (including each other), can provide contextually-relevant and structured data for enabling the construction of such models. Then, a possible platform technology for creating such sensors is examined, namely Large-Area Electronics (LAE). The challenges of realizing full systems from this are explored. In particular, perceptive systems present demanding functional requirements, but, through emerging algorithms from statistical signal processing and machine learning, also open up new opportunities for addressing technological limitations. Several LAE systems for human monitoring are presented, demonstrating the potentials.