possible to be transmitted depends on the selected MCS and the transmission rank, showing that HD could perform better than FD.
Given the conclusions of the limited benefits provided by FD in indoor small cell networks, other applications where this technology can provide higher benefits is studied. In particular, the potential of FD in reducing the discovery time in D2D communication. This type of communication allows devices to communicate directly among them, thus offloading the infrastruc-ture. The discovery mechanism is required to detect peer devices in the sur-roundings, and should be completed within 10 milliseconds. To achieve such strict control latency requirement, FD technology is considered, since it over-comes the HD constraint of not being able to listen to neighboring transmis-sions while transmitting the discovery message.
The analysis shows that FD has potential in reducing the discovery time if interference cancellation receivers are assumed. Results indicate that the optimal transmission probability leading to the minimum discovery time is different depending on the scenario, e.g., number of neighbors. In addition, interference management techniques are required to avoid increasing the la-tency. The proposed solution includes the dynamic selection of the transmis-sion probability based on an estimation of the number of neighboring devices and a signaling scheme to reduce the network interference. Simulations show that the proposed strategy reduces the discovery time and can meet the la-tency target defined for 5G at the expenses of a minor increase in control overhead.
2 Future Work
The work presented in this dissertation provides many possibilities for future research on small cell networks and applications where full duplex technol-ogy may bring large benefits.
When dynamic TDD is considered to improve the capacity of 5G small cells, mechanisms to stabilize the interference are required. Part II of the the-sis describes interference management techniques as a solution to deal with the drawback introduced by dynamic TDD. The approach of using coordi-nation by means of a centralized controller could be evaluated, for exam-ple following the coordinated multi-point princiexam-ple, where users receive the data from multiple base stations to improve the signal quality. The design should be carefully addressed, since the generated control overhead and the increased delay may bring more harm than benefits.
The study of FD in 5G indoor small cells led to the conclusion that un-der strong interference, HD may outperform FD. Consequently, the design of a hybrid HD/FD scheduler that chooses the most appropriate transmission
Conclusion
mode depending on the interference conditions could benefit the system in terms of throughput and delay. An analysis of the SINR range where each transmission mode brings higher benefits to the system can be carried out.
The studies involving FD were performed assuming ideal SI cancellation.
Therefore, an interesting work could be to quantify the impact of non-ideal SI cancellation. A model which depends on the transmit power and the band-width could be derived and evaluated. In addition, 5G is targeted to be cost-effective. Therefore, the model could also contemplate simpler cancella-tion techniques and study its impact on the FD performance.
A third FD application could be studied: the BS acting as a FD relay, i.e., the BS forwarding data from one user to another one. This case faces simi-lar conditions as the BS FD case, since intra-cell interference is also present.
However, from a traffic perspective, it is required that two users need to ex-change data. There are two options on implementing such a relay. The first one is to allow the BS to decode the data and store it until the receiving node acknowledges it. This approach would improve the throughput but not the delay. The second options is to forward the data as it arrives at the BS, with-out performing error check. The main drawback of this approach is that in case of link failure, both UL and DL transmissions need to be rescheduled.
On the other hand, the delay can be improved if the decoding is performed successfully at the first transmission attempt.
The indoor small cell network targeted in this work is an interference limited scenario. Therefore, FD could be evaluated in other scenarios, e.g., a macro cell. In this case, since the transmit power of the base station is higher than in the indoor scenario, a model to take into account the impact of non-ideal SI cancellation should be included. Interference management and power control techniques would also be required to limit the interfer-ence among base stations. Furthermore, a smart scheduler that selects the pair of users which reduces the harm from the UL to the DL user would also be needed.
In terms of D2D communication, future work could focus on the design of the discovery message and how feedback is piggybacked: which is the maximum number of acknowledgments that could be embedded in the dis-covery message, which is the information required to be transmitted or how to identify each device. In this work, an assumption for the discovery mes-sage transmission was to use a fixed MCS so no link adaptation was required and the control overhead could be reduced. However, it could be interest-ing to consider link adaptation techniques and compare both approaches, in terms of control overhead, robustness and discovery time.
The proposed FD solution of adapting dynamically the transmission prob-ability and signaling such parameter could be evaluated in scenarios with mobility such as V2X. This scenario is challenging since the number of neigh-bors varies rapidly and often.
References
References
[1] P. Mogensenet al., “Centimeter-wave concept for 5G ultra-dense small cells,” in IEEE 79th Vehicular Technology Conference (VTC Spring), May 2014.
[2] D. Catania, “Performance of 5G small cells using flexible TDD,” Ph.D. disserta-tion, Department of Electronic Systems, Aalborg University, October 2015.
[3] F. M. L. Tavares, G. Berardinelli, N. H. Mahmood, D. Catania, T. B. Sørensen, and P. E. Mogensen, “Interference-robust air interface for 5G ultra-dense small cells,”
Journal of Signal Processing Systems, pp. 1–14, 5 2016.
[4] M. Heinoet al., “Recent advances in antenna design and interference cancellation algorithms for in-band full duplex relays,”IEEE Communications Magazine, vol. 53, no. 5, pp. 91–101, 2015.
References