Cheat-Proof Power Control for Full-Duplex Small Cells: Towards 5G

The demand for high speed data will be increased ever than before in fifth generation (5G) of wireless communication networks [1]. One of the most promising ways to increase data rate is to make the wireless cell smaller in size and densify the network with a large number of low power base stations (i.e., Small Cells). These small cells can increase the wireless data rate by providing higher quality links between the transmitters and the receivers and also by exploiting more spatial spectrum reuse. Full-duplex transmission (i.e., transmitting and receiving at the same time in same frequency band) is another recently emerging technology which can theoretically double the date rate of a wireless network [2]. Recent studies have shown that full-duplex technology works better for low power transmission nodes [2]. Therefore, exploiting the possibilities to deploy full-duplex small cells is crucial for future 5G networks. As 5G networks will be highly dense with different types of wireless nodes [3], centralized resource allocation will need an immense amount of information exchange. However, the backhaul capacity available for information exchange among small cells is still limited. Hence, centralized network control will be highly inefficient and expensive. Therefore, it is also essential to develop distributed resource allocation techniques which reduce the information exchange among the network nodes. Moreover, most of the existing distributed power control algorithms fail when transmitters can cheat and transmit at their maximum possible power. This research addresses the problem of distributed cheat-proof power control for full-duplex small cell networks. We model the network as a non-cooperative repeated game with imperfect information and propose a distributed algorithm to achieve a public perfect equilibrium (PPE) which is also Pareto optimal. The proposed algorithm can also prevent cheating. Existence of PPE and the convergence of the algorithm to a Pareto optimal solution is proven analytically. Numerical results are presented to show the effectiveness of the algorithm.