Seokki Kim

Asymptotically Optimal Cooperative Jamming for Physical Layer Security

Design of effective cooperative jamming (CJ) algorithm is studied in this paper to maximize the achievable secrecy rate when the total transmit power of the source and multiple trusted terminals is constrained. Recently, the same problem was studied in [1] and an optimal algorithm was proposed involving a one-dimensional exhaustive searching. However, the computational complexity of such exhaustive searching could be very high, which may limit the practical use of the optimal algorithm. We propose an asymptotically optimal algorithm, involving only a fast line searching, which can guarantee to achieve the global optimality when the total transmit power goes to infinity. Numerical results demonstrate that the proposed asymptotically optimal algorithm essentially gives the same performance as the algorithm in [1, (44)] but with much lower computational complexity.

Relay selection in multiple clustered relay networks

We consider the downlink of a realistic cellular network composed of multiple clusters, each of which includes multiple relays and a destination terminal. In this scenario, the interference among the clusters is a critical issue, affecting the end-to-end performance of the system significantly. Considering the inter-cluster interference, we propose optimal and suboptimal relay selection schemes to maximize the sum rate of the network. The first algorithm is a truly optimal network-wise relay selection scheme which provides the ultimate performance upper bound. However, the computational complexity and the signaling overhead involved with the optimal scheme are fairly high. Addressing the issues, we propose a suboptimal cluster-wise scheme with a much less computational complexity and signaling overhead. Numerical results demonstrate that the proposed algorithms provide excellent performance.

Effect of narrow band signal interference caused by CFO on OFDM system for proximity-based direct communication

Proximity-based direct communication is the technology that devices in the near distance can use the radio resource efficiently by operating distributedly without control of BS or AP. Many technologies are proposed for proximity-based direct communication and taking the operation in unlicensed band into account. Thus, coexistence techniques are needed for coexistence with different kind of devices which operate in unlicensed band such as Wireless LAN. A blocking signal transmission scheme to avoid contention with other devices by transmitting narrow band signal is proposed as one of the coexistence schemes for coexistence with different kind of devices. In this work, we present analysis of the effect of narrow band interference such as blocking signal caused by CFO on OFDM system for proximity-based direct communication and validate the theoretical results by means of computer simulation.

System Level Simulation of mmWave Based Mobile Xhaul Networks

To support tens of Giga-bits/second (Gbps) data rate, 5G mobile communication systems consider the integration of backhaul and fronthaul links into Xhaul networks. This paper introduces the system-level simulation for mmWave based mobile Xhaul networks. Network scenarios, hybrid beamforming, with large antenna elements, and link-level models are discussed based on the results of the system-level simulation for the mobile Xhaul network. System-level simulation results show that the mobile Xhau network using a 40 GHz carrier frequency band can support a 20 Gbps data rate.

Interference-Aware Clustering Algorithms in Multi-relay Cellular Networks

We consider the downlink of a wireless cellular network which consists of one Macro Base Station (MBS), multiple relays and User Equipments (UEs). In this scenario, we propose relay clustering algorithms to address the problem of inter-cluster interference. In each cluster, we adopt a time-division relaying scheme to avoid intra-cluster interference, but the inter-cluster interference still remains. In interference-aware network, an optimal clustering algorithm requires an exhaustive search, which is infeasible. Hence, we propose a clustering algorithm which improves the system capacity by exploiting the inter-cluster information. But this algorithm is implemented in a centralized manner with high complexity that requires the calculation of inter-cluster information. To circumvent this, we propose a cluster-wise distributed algorithm with manageable complexity, which shows the performance within acceptable range from the optimal one, far better than simple and intuitive algorithms because of the effect of inter-cluster interference.