Providing seamless, high quality wireless service anytime and anywhere requires substantial structural changes in today's macro-cellular networks. One such change, introducing small cell base stations, is seen as a highly promising solution. However, it requires meeting fundamental challenges: 1) nodes? self-organization, 2) network heterogeneity, and 3) high sensitivity of resource allocation to the system parameters. The proposed research addresses these challenges by exploring a dimension that has often been overlooked: the user's context. To achieve this goal, first, machine learning techniques are proposed to extract context from three dimensions: device, geo-location, and social metrics. Then, context-aware resource management schemes are developed by advancing novel techniques from matching theory - a powerful tool from economics and game theory. Subsequently, the learned context is leveraged to devise cooperative small cell models using new tools from coalitional game theory. Comprehensive evaluation is done via testbed implementation and software simulations.
The developed analytical tools will lay the foundations of context-aware, self-organizing small cell networks and will impact multiple disciplines such as communications, game theory, and social sciences. The generated results will provide fresh ideas for developing new small cell products. The research is fully integrated into the educational plan via incorporation in new and existing courses as well as training students via mentoring, participation in testbed development, and internships at industrial labs. A developed small cell educational tool will foster this integration via new hands-on activities and demonstrations to the community. Specialized outreach activities will contribute to increasing the participation of minority high school students in science and engineering.
