NeTS: Small: Collaborative Research: Advanced Algorithmic Tools for Discovery in Cognitive Radio Networks

Northeastern University, Boston MA

Investigators

Abstract

The prevalence of embedded computing and wireless technology has given rise to remarkable new opportunities for technology to provide and improve safety and services. Such applications include on-the-fly traffic management to improve safety and reduce congestion, establishing ad-hoc networks for disaster recovery, and maintaining sensor networks for military or scientific purposes. In all these cases, wireless devices must discover each other's existence, configure their wireless radios to maximize efficiency, and conserve battery power. This project will develop new methods to address these fundamental problems of discovery and efficiency. Additionally, the project will develop and new techniques and tools for evaluating the efficacy of proposed solutions. The project also incorporates development of undergraduate- and graduate-level courses, outreach via educational opportunities for high-school teachers and students, expanding participation of undergraduates and underrepresented groups in research, and support for female students in computing. This project aims to develop and analyze advanced algorithmic tools for spatio-temporal peer discovery, a fundamental primitive for ad-hoc networks. The proposal focuses on cognitive radio networks characterized by high mobility, congested spectrum, motion-induced channel outages and energy-constrained commodity devices. The development of efficient discovery schemes with both provable guarantees and strong practical performance will be critical for the build out of such networks for large scale applications such as road safety and emergency response for driver-less fleets. The project has four principal thrusts: (1) Multi-channel discovery, where nodes equipped with spectrally agile radios channel hop to achieve rendezvous using deterministic and probabilistic techniques for achieving multi-channel discovery, drawing inspiration from constructive Ramsey theory, cyclic algebraic geometry codes, number theoretic sieves, and a novel pseudo-random construction: the synchronizing sampler; (2) Energy-constrained single channel discovery algorithms where mobile nodes duty cycle to save battery, via connections with additive combinatorics, algebraic geometry, and expander graphs; (3) Spatial discovery through signal strength by applying techniques from computational and convex geometry for estimations of trajectories within vehicular environments; and (4) Validation via a simulation environment using the ns-3 simulator with appropriate extensions for multi-channel vehicular networks. While focusing on vehicular networks, these techniques can be applied in a wide variety of different networks, such as medical and telemetry networks as well.