Secrecy Enhancement in AF Relaying Assisted SWIPT-NOMA via Power Optimization and Control-Jamming

Abstract

This work aims at achieving the secrecy rate (SR) enhancement in amplify-and-forward (AF) relay-aided simultaneous wireless information and power transfer (SWIPT)-NOMA network via control-jamming (CJ) whilst satisfying the constraints on total transmit power budget. Here, in order to coordinate the information processing and energy-harvesting (EH), power-splitting (PS) protocol is deployed at the relay. For comparison purposes, in addition to CJ, the SR of with-jamming (J) and without-jamming (WJ) conditions are also highlighted in the existence of an eavesdropper (Eve). Moreover, the analytical expression for Ergodic SR is evaluated at the NOMA users under the CJ scenario. Next, to maximize the SR of the users further, an optimal power allocation (OPA) scheme has been adopted, which involves both particle swarm optimization (PSO) and Lagrangian multiplier (LM)-based approaches to acquire the optimized power factors of various nodes of the network. In addition, we analyzed the affect of major parameters such as: $J \to Eve$ distance, transmit signal-to-noise ratio (SNR), and $R \to Eve$ distance on the SR by exploiting distinct jamming scenarios. Finally, a significant secrecy performance gain in terms of SR is verified with proposed CJ over other jamming conditions and OPA over the fixed power allocation (FPA) and it is confirmed through Monte-Carlo based simulations.

Appendices

Appendix A

Evaluation of SR in WJ condition

The expression for SR at the NOMA users in WJ (absence of jammer) condition is formulated as

$$ C_{AF\_Ui}^{wj} = 0.5*\log_{2} (1 + \gamma_{Ui}^{wj} ) - 0.5*\log_{2} (1 + \gamma_{Ei}^{wj} )\quad i = 1,2 \, $$

(A.1)

In this scenario, due to the absence of jammer, the Eve acquires only the confidential information via the EH relay and it can be designated as

$$ y_{e} = y_{re} = kh_{re} \left( {h_{sr} (\sqrt {(1 - \varphi )P_{s} a_{1} } x_{1} + \sqrt {(1 - \varphi )P_{s} a_{2} } x_{2} ) + n_{sr} } \right) + n_{re} $$

(A.2)

So, the SNR of the Eve is determined as

$$ \gamma_{Ei}^{wj} = \frac{{k^{2} (1 - \varphi )P_{s} a_{i} \left| {h_{sr} } \right|^{2} \left| {h_{re} } \right|^{2} }}{{(k^{2} \left| {h_{re} } \right|^{2} + 1)N_{0} }}\quad i = 1,2 $$

(A.3)

Finally, we substitute (A.3), (8), and (9) in (A.1) to evaluate the SR at the NOMA users.

Appendix B

Evaluation of SR in ‘J’ condition

The expression for SR at the NOMA users in ‘J’ condition is denoted as

$$ C_{AF\_Ui}^{j} = 0.5*\log_{2} (1 + \gamma_{Ui}^{j} ) - 0.5*\log_{2} (1 + \gamma_{Ei}^{j} )\quad i = 1,2 \, $$

(A.4)

In this scenario, users acquire confidential message from the relay and also the jamming signal from the jammer, which can be denoted as

$$ y_{{u_{i} }} = y_{{ju_{i} }} + \, y_{{ru_{i} }} $$

(A.5)

where

$$ y_{{ju_{i} }} = \sqrt {P_{j} } \, h_{{ju_{i} }} x + n_{{ju_{i} }} \& y_{{ru_{i} }} = kh_{{ru_{i} }} \left( {h_{sr} (\sqrt {(1 - \varphi )P_{s} a_{1} } x_{1} + \sqrt {(1 - \varphi )P_{s} a_{2} } x_{2} ) + n_{sr} } \right) + n_{{ru_{i} }} \quad i = 1,2 $$

(A.6)

From (A.6), the obtained SINR and SNR for the users U₁ and U₂ to detect x₁ and x₂ are respectively given as

$$ \gamma_{U1}^{j} = \frac{{k^{2} (1 - \varphi )P_{s} a_{1} \left| {h_{sr} } \right|^{2} \left| {h_{ru1} } \right|^{2} }}{{k^{2} (1 - \varphi )P_{s} a_{2} \left| {h_{sr} } \right|^{2} \left| {h_{ru1} } \right|^{2} + (k^{2} \left| {h_{ru1} } \right|^{2} + 2)N_{0} + P_{j} \left| {h_{{ju_{1} }} } \right|^{2} }}\,\& \,\gamma_{U2}^{j} = \frac{{k^{2} (1 - \varphi )P_{s} a_{2} \left| {h_{sr} } \right|^{2} \left| {h_{ru2} } \right|^{2} }}{{(\beta^{2} \left| {h_{ru2} } \right|^{2} + 2)N_{0} + P_{j} \left| {h_{{ju_{2} }} } \right|^{2} }} $$

(A.7)

The SNR obtained at the Eve is given in (11).

SR can be evaluated by substituting (A.7) and (11) in (A.4).