This project addresses fundamental issues in the White-Cover network which provides data coverage services in the densely populated areas with the TV White Space spectrum. It is motivated by the fact that a White-Cover cell will likely host hundreds of users while supporting highly dynamic data traffic, a scenario beyond the capabilities of existing network protocols. In addition, the White-Cover devices will likely have to coexist in the same channel with heterogeneous devices due to the lack of free TV channels in the densely populated areas. These challenges are addressed by a combination of novel physical layer techniques and advanced machine learning techniques, and the solutions are enhanced by further optimizations. This project studies: 1) efficient physical layer schemes for node state detection with which the transmissions in the cell can be scheduled very efficiently, 2) advanced protocols that adapt to the heterogamous interferences conditions with the reinforcement learning technique, 3) optimized protocol for power consumption and integrated partial packet recovery, and 4) network-wide performance analysis and planning.
The White-Cover network designed in this project provides a better alternative to the current technologies for the TV White Space. Users may enjoy data coverage at much less cost than from the cellular phone networks and new business may be enabled. The expected results of this project include the protocols and algorithms optimized for the unique features of White-Cover and tested with experiments. The results will be disseminated in conferences and journals; the source code will be freely available online.
