CNS Core: Small: Sensor-based Wireless Network Management via Edge Computing

University Of Colorado At Boulder, Boulder CO

Investigators

Abstract

As Internet-of-Things, mobile health, autonomous vehicles, cyber-physical systems, virtual reality, and smart environments proliferate, wireless spectrum demand continues to grow at a significant rate. This proposal investigates new ways of managing wireless networks through the advanced integration of emerging trends in software-driven networking, ubiquitous computing deployments, and sensor-driven analysis. The proposal supports the National Science Foundation's mission to advance prosperity by enabling wireless industries to more efficiently and flexibly control their networks. The investigator will work closely with the wireless communication industry that contributes $475 billion annually to America's economy each year and supports up to 4.7 million jobs nationally. The project further benefits society by developing sensor-based computing educational modules that will be released on TeachEngineering, a free searchable engineering curricula library for K-12 educators. Two major thrusts will be studied to build an architecture that enables practical sensor-based wireless network management. The first thrust aims to utilize sensor data to more effectively and flexibly manage the wireless network. First, objects in the sensory domain will be aligned with objects in the wireless domain through behavior analysis techniques in both domains. Second, techniques will be designed to flexibly schedule network traffic based on the actions of mobile device users. Third, a programmable framework and corresponding feasibility tools allow administrators to configure this new type of network architecture. A second major thrust in this proposal makes analyzing sensory data at the network edge more efficient. First, user and traffic context are used to determine how much edge resources should be allocated for sensor analysis. Second, computational demands on network edge resources are lessened via context-aware sampling rate algorithms. In addition, new techniques utilize wireless information to lessen sensor-based processing. The research results will be incorporated into courses at the University of Colorado Boulder, scientific publications, and released as open-source software. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.