PaperlayerView exact revision on arXiv

Source-linked AI summary

IRS-based Wireless Jamming Attacks: When Jammers can Attack without Power

Bin Lyu, Dinh Thai Hoang, Shimin Gong, Dusit Niyato, Dong In Kim

arXiv:2001.01887v2cs.ITeess.SPmath.IT

TL;DR

The paper considers how an IRS, which can harm legitimate communications, can act as a jammer without generating signals using its own energy. It jointly designs reflection magnitudes and discrete phase shifts to reduce received signal power. Simulations report about 22 dB SINR degradation versus no jamming and better performance than a 15 dBm active jammer in the stated comparison.

  • Problem

    IRS deployment can cause harmful interference to critical wireless communications, motivating study of its use to degrade the legitimate receiver’s SINR.

    Source: §I · p. 2

  • Method

    The IRS jammer reflects the legitimate transmitter’s signals and jointly optimizes reflection magnitudes and discrete phase shifts to minimize received power at the LR.

    Source: §I · p. 3

  • Results

    About 22 dB SINR degradation versus no jamming; the IRS jammer also outperforms the active jammer at 15 dBm in the reported comparison.

    Source: §IV · pp. 9–10

  • Takeaways & Limitations

    The IRS jammer can reduce LR SINR without using internal energy to generate jamming signals and can outperform conventional active jamming in some scenarios.

    Source: §V · pp. 11–12

Abstract

from arXiv · show

This paper proposes to use Intelligent Reflecting Surface (IRS) as a green jammer to attack a legitimate communication without using any internal energy to generate jamming signals. In particular, the IRS is used to intelligently reflect the signals from the legitimate transmitter to the legitimate receiver (LR) to guarantee that the received signals from direct and reflecting links can be added destructively, which thus diminishes the Signal-to-Interference-plus-Noise Ratio (SINR) at the LR. To minimize the received signal power at the LR, we consider the joint optimization of magnitudes of reflection coefficients and discrete phase shifts at the IRS. Based on the block coordinate descent, semidefinite relaxation, and Gaussian randomization techniques, the solution can be obtained efficiently. Through simulation results, we show that by using the IRS-based jammer, we can reduce the signal power received at the LR by up to 99\%. Interestingly, the performance of the proposed IRS-based jammer is even better than that of the conventional active jamming attacks in some scenarios.

I. INTRODUCTION

IRS can improve wireless links through controlled reflections, but it can also be used to interfere with legitimate communications. The paper proposes an IRS jammer that redirects the victim’s own signal and jointly designs reflection magnitudes and phases to reduce received power.

I. INTRODUCTION

  • An IRS uses independently controlled passive elements to adjust reflected-signal amplitudes and phases, steering reflected electromagnetic waves to affect received power.The IRS is presented as a means of improving spectrum and energy efficiency in future wireless systems.Source: §I · pp. 1–2
  • Unlike prior IRS studies focused on improving communications or secrecy, this work examines using an IRS to degrade the legitimate receiver’s SINR.The cited prior work considered phase-shift design with reflection coefficients set to one.Source: §I · p. 2
  • The proposed attack reflects the victim communication’s own signals, avoiding internal energy for signal generation and making detection and prevention difficult.The reflection coefficients and phase shifts are changed to carry out the attack.Source: §I · p. 3
  • The proposed IRS jammer reduces received signal power at the LR by up to 99% and can outperform conventional active jamming in some scenarios.It uses the legitimate transmitter’s signal rather than internal energy to generate jamming signals.Source: §I · p. 3
  • The jammer jointly optimizes reflection magnitudes and discrete phase shifts, using alternating BCD with SDR, Gaussian randomization, and phase quantization for the phase-design subproblem.The paper uses CVX to solve the other subproblem.Source: §I · p. 3

II. SYSTEM MODEL

The system places an IRS-based jammer between a multi-antenna legitimate transmitter and a single-antenna receiver. Its passive elements control reflection amplitudes and phases, with discrete phase choices and only first reflections included in the model.

II. SYSTEM MODEL

  • The model has an LT with M antennas, an LR with one antenna, and an IRS jammer equipped with N passive reflecting elements.Each IRS element has adjustable amplitude and phase and reflects incident signals independently.Source: §II · pp. 3–4
  • Each element’s reflection coefficient has magnitude β_n in [0, 1] and a phase shift θ_n chosen from uniformly quantized levels over [0, 2π).The number of phase levels is L = 2^b, where b is the phase resolution in bits.Source: §II · p. 4
  • The LT uses linear beamforming with total transmit power P_T, and the model ignores IRS reflections after the first bounce because of substantial path loss.The received signal includes additive white Gaussian noise at the LR.Source: §II · p. 4

A. Problem Formulation

The paper formulates IRS jamming as minimizing received signal power at the LR by jointly choosing reflection magnitudes and phase shifts. The resulting problem is difficult because the variables are coupled and phase shifts are discrete.

A. Problem Formulation

  • The optimization objective is to minimize received signal power at the LR by jointly selecting reflection magnitudes and phase shifts.The design variables are denoted β and θ.Source: A. Problem Formulation · p. 5
  • The objective is non-convex because the reflection magnitudes β and phase shifts θ are coupled.Source: A. Problem Formulation · p. 5
  • Discrete phase-shift constraints make P1 a mixed-integer nonlinear program that is typically NP-hard and difficult to solve directly.Source: A. Problem Formulation · p. 5

B. Proposed Solution

The proposed solution uses block coordinate descent to optimize reflection magnitudes and phase shifts in alternating steps.

B. Proposed Solution

1) Phase shift design:

For fixed reflection magnitudes, the phase-shift design minimizes received signal power while accommodating discrete phase choices. It uses semidefinite relaxation and Gaussian randomization to obtain a quantized approximate solution.

1) Phase shift design:

  • For fixed reflection magnitudes, the phase-shift subproblem relaxes discrete phases to continuous values before solving.The final phase shifts are quantized to their nearest permitted discrete values.Source: 1) Phase shift design · pp. 5, 7
  • The formulation assumes each element’s reflection magnitude is independent of its phase shift; a practical nonlinear dependence is left for future investigation.The assumed ideal phase-shift model is stated to be for simplicity.Source: 1) Phase shift design · p. 5
  • The unit-modulus phase vector is lifted to V = μμH, yielding a reformulated objective with diagonal-one and rank-one constraints.The formulation also requires V ⪰ 0.Source: 1) Phase shift design · pp. 5–6
  • Semidefinite relaxation removes the rank-one constraint, and Gaussian randomization constructs approximate solutions from the relaxed SDP.The relaxed solution generally is not rank one and gives a lower bound on received signal power.Source: 1) Phase shift design · p. 6
  • Among generated candidates, the method selects the quantized phase shifts that minimize received signal power at the legitimate receiver.The paper describes this as a high-accuracy sub-optimal solution.Source: 1) Phase shift design · p. 7

2) Reflection coefficient design:

With phase shifts fixed, the method designs reflection magnitudes through a convex reformulation. It alternates phase and magnitude updates, with received signal power non-increasing across iterations until the specified threshold is reached.

2) Reflection coefficient design:

  • Given phase shifts, reflection magnitudes are designed by solving P3, which is reformulated as the convex optimization problem P3.1.The paper states that P3.1 can be solved directly with CVX.Source: 2) Reflection coefficient design · pp. 7–8
  • Received signal power is non-increasing after sequential phase and magnitude updates, and Algorithm 2 converges to the given threshold..Source: 2) Reflection coefficient design · p. 8
  • The joint algorithm alternates between solving the phase subproblem for fixed magnitudes and updating magnitudes for the resulting phase shifts.The listed steps continue until convergence is achieved.Source: 2) Reflection coefficient design · p. 8

IV. PERFORMANCE EVALUATION

The evaluation compares the IRS jammer with active jamming and no jamming while varying transmitter power, reflecting-element count, and IRS placement. Its relative performance depends on the scenario, including the distances between the IRS and legitimate devices.

IV. PERFORMANCE EVALUATION

  • The simulations use a 2D topology and compare against active jamming and a no-jamming scheme; the algorithm’s P2 and P3.1 subproblems are solved once sequentially.The stated default settings include PT = 30 dBm, M = 8, N = 150, and b = 5.Source: §IV · p. 8
  • About 22 dB lower SINR at the LR is achieved by the proposed scheme compared with no jamming.The comparison varies the LT’s transmit power; at Pa = 15 dBm, the IRS jammer also outperforms active jamming.Source: Figure 2 · p. 9, §IV · pp. 9–10
  • With 200 reflecting elements, the IRS jammer can outperform active jamming with Pa = 35 dBm.The attack performance improves as the number of reflecting elements increases.Source: Figure 3 · p. 9, §IV · p. 10
  • For LT–IRS distances shorter than 7 m, the proposed IRS-jamming scheme performs better than active jamming at Pa = 15 dBm; beyond 7 m, it performs less well than that scheme.As LT–IRS distance increases, received signal power from reflecting links falls, limiting destructive addition and reducing IRS-jamming effectiveness.Source: §IV · p. 11
  • When LR–IRS distance is below 8 m, the IRS jammer outperforms active jamming with Pa = 15 dBm; at 5 m, it approaches active jamming with Pa = 35 dBm.The LT–IRS distance is held at 5 m.Source: Figure 5 · p. 11, §IV · p. 11

V. CONCLUSIONS

The proposed IRS jammer reuses the legitimate transmitter’s signals and controls its reflections to reduce the legitimate receiver’s SINR. Simulations show it can outperform conventional active jamming in some scenarios.

  • Unlike conventional wireless jammers, it uses the LT’s signals rather than its own energy to reduce the LR’s SINR.Source: §V · p. 11
  • By controlling reflection coefficients and phase shifts, the IRS can reflect signals that significantly reduce received signal power at the LR.Source: §V · pp. 11–12
  • The IRS jammer can outperform conventional active jamming in some scenarios while reducing the LR’s SINR.Source: §V · p. 12
Loading 2001.01887v2…