| Summary: | 碩士 === 國立清華大學 === 通訊工程研究所 === 98 === In this thesis, we consider a multiple access wiretap channel where multiple transmitters want to communicate with a receiver (e.g., base station) in the presence of eavesdroppers. The eavesdroppers aim to eavesdrop the secret messages sent by the transmitters.
We consider the so called achievable secrecy sum rate, which is the achievable sum rate under which the receiver can reliably decode the message whereas all the eavesdroppers cannot retrieve any information. We focus on the achievable secrecy sum rate maximization problem where the transmit signal covariance matrices of transmitters are optimized subject to individual power constraints. This class of optimization problems is nonconvex; therefore, it’s difficult to obtain its global optimal solution in general. In this thesis, we study three different scenarios, namely, the MISOSE (multiple-input, single-output, single eavesdropper-antenna), the MISOME (multiple-input, single-output, multiple eavesdropper-antennas) and the MIMOSE (multiple-input, multiple-output, single eavesdropper-antenna). For each scenario, we investigate efficient methods for handling the associated optimization problem. We will show that, when each of the eavesdroppers has a single antenna, the secrecy sum rate maximization problem can be reformulated, and solved efficiently. However, when each eavesdropper has multiple antennas, we can only obtain suboptimal solutions.
Extensive simulation results are presented to demonstrate the effectiveness of the proposed methods. These results will show how the number of antennas, number of transmitters, and number of eavesdroppers affect the achievable secrecy sum rate.
|
|---|
