Cooperative Pilot Spoofing in MU-MIMO Systems

TL;DR

This paper introduces a cooperative pilot spoofing attack in MU-MIMO systems that significantly degrades downlink throughput by impairing channel estimation, with an analytical framework for optimal attack strategies and performance evaluation.

Contribution

It proposes a novel cooperative attack method in MU-MIMO systems, analyzes its impact on channel estimation and throughput, and derives an optimal attack strategy with convex optimization.

Findings

CPSA severely reduces downlink sum-rate as attacker power increases

The attack's impact intensifies with more attackers involved

Optimal attack strategy can be efficiently computed using convex optimization

Abstract

In this letter, we consider downlink transmission of a multiuser multiple-input multiple-output (MU-MIMO) system with zero-forcing (ZF) precoders in the presence of multiple attackers. We propose a cooperative pilot spoofing attack (CPSA), where the attackers collaboratively impair the channel estimations in the uplink channel training phase, aiming at deteriorating the downlink throughput of the whole cell. We first evaluate the impacts of CPSA on the channel estimation and the downlink ZF precoding design, and then we derive an analytical expression for the achievable downlink sum-rate. Furthermore, we investigate the optimal attack strategy to minimize the achievable downlink sum-rate. We show that the optimization problem under consideration is a convex one so the global optimum could be obtained conveniently. Numerical results show that the CPSA attack results in a severe performance deterioration with the increase in the attacking power and the number of attackers.