NeTS: Small: A Theoretical Approach to MAC Design for Communication Between Low Cost, Ultra-Low Power Devices

Columbia University, New York NY

Investigators

Abstract

The Internet-of-Things is being extended to everyday physical objects by placing tags on the objects, which have the ability to sense physical attributes of the environment and, along with the object's identity, relay information for processing. The current convention is to have passive tags that must communicate directly with a powerful tag-reading device, but this limits the tags' ability to convey information to scenarios where the readers are sufficiently prevalent. An alternative, more effective paradigm, is to allow the tags to exchange information with one another, passing information tag-to-tag until reaching a tag that is in range with a gateway (reader) device. Under this latter paradigm, the likelihood of objects being "disconnected" from the network will significantly reduce, making tracking and monitoring more ubiquitous and complete. Misplacing objects will become a problem of the past, especially in regions where infrastructure is difficult to deploy, e.g., in emergency and triage settings, which often have the major challenge of finding objects or people or assessing conditions. This project will have broad impact, bringing significant benefits to users and society at large. Specifically, the proposed research will impact society in facilitating the path toward extending the realm of the Internet-of-Things to everyday objects; instances include tracking of everyday objects, as well as facilitating the tracking of objects in specialized domains such as medicine, construction, and shipping. The educational activities will create a rich set of opportunities for students to involve themselves in algorithms, modeling, and evaluation work. The research will also contribute to the educational component as the principal investigators (PIs) have and will continue to design courses around this emerging topic and will continue to involve not only PhD students, but also high school, undergraduate, and masters students in research that directly relates to this proposal. The PIs explore, from a theoretical perspective, the fundamental limits of tag-to-tag communication, in particular how to maximize communication rates in networks built from these ultra-low power devices in environments where they must sleep for large fractions of time (in order to harvest/conserve energy) and operate in an unsynchronized manner. From a practical perspective, protocols will be designed and tested that can perform transmission at rates close to these fundamental limits. The research will create a stochastic optimization framework that is used in design and development of practical distributed protocols. The theoretical algorithms obtained in this research will be also tested and implemented using actual prototypes.