Joint Jammer Mitigation and Data Detection

TL;DR

This work introduces Joint Jammer Mitigation and Data Detection (JMD), a MIMO receiver paradigm that eliminates the need for dedicated jammer training by jointly estimating the jammer interference subspace and data over a coherence frame. It develops two algorithms, SANDMAN (linear-channel-estimation) and MAED (joint-channel-estimation), that solve nonconvex optimization via alternating minimization, a convex-hull relaxation of the symbol set, and a constellation-promoting prior, achieving jammer mitigation without sacrificing data rate. Theoretical results characterize eclipsing probabilities and provide guarantees under zero thermal noise, while simulations across diverse jammers (smart, distributed, multi-antenna, and dynamic) show JMD matching genie-assisted baselines and outperforming training-based schemes. The work highlights the practical potential of JMD for robust, high-rate wireless communications in adversarial environments, while noting current limitations and directions for enhancement such as universal transforms to further reduce complexity and improve performance in sparse-jammer regimes.

Abstract

Multi-antenna (or MIMO) processing is a promising solution to the problem of jammer mitigation. Existing methods mitigate the jammer based on an estimate of its spatial signature that is acquired through a dedicated training phase. This strategy has two main drawbacks: (i) it reduces the communication rate since no data can be transmitted during the training phase and (ii) it can be evaded by smart or multi-antenna jammers that do not transmit during the training phase or that dynamically change their subspace through time-varying beamforming. To address these drawbacks, we propose Joint jammer Mitigation and data Detection (JMD), a novel paradigm for MIMO jammer mitigation. The core idea of JMD is to estimate and remove the jammer interference subspace jointly with detecting the legitimate transmit data over multiple time slots. Doing so removes the need for a dedicated and rate-reducing training period while being able to mitigate smart and dynamic multi-antenna jammers. We provide two JMD-type algorithms, SANDMAN and MAED, that differ in the way they estimate the channels of the legitimate transmitters and achieve different complexity-performance tradeoffs. Extensive simulations demonstrate the efficacy of JMD for jammer mitigation.

Figures (4)

• Figure 1: The bound $p_e$ (and its approximation $\tilde{p}_e$) on the probability that a single-antenna jammer eclipses vs. the number of jammed symbols $\|\mathbf{w}\xspace\|_0$, for different numbers $U$ of UEs. The bound on the probability of eclipsing decreases exponentially in the number of jammed symbols.
• Figure 2: Trade-off between the relative achievable rate $r$ and the lowest SNR for which the different receivers satisfy the criterion $\text{MER}\leq17.5\%$ when mitigating a single-antenna barrage jammer [inner color=white, fill color=mylightgray1, outer color=mylightgray1]1.
• Figure 3: Uncoded bit error-rate (BER) vs. SNR performance of different receivers when mitigating different kinds of jammers, including smart, distributed, and dynamic multi-antenna jammers.
• Figure 4: Illustration of $\mathcal{S}\xspace$ and $\Delta\mathcal{S}\xspace$. Values in $\Delta\mathcal{S}\xspace$ which are contained in $\mathfrak{e}_k(s_k)$ with probability $1$ are circumscribed with a drawn red line; values that are contained with probability $\frac{1}{2}$ are circumscribed with a dashed red line; values that are contained with probability $\frac{1}{4}$ are circumscribed with a dotted red line.

Theorems & Definitions (21)

• Definition 1: Eclipsing with perfect CSI
• Theorem 1
• Theorem 2
• Definition 2: Eclipsing with channel estimation
• Theorem 3
• Theorem 4
• Theorem 5
• Theorem 6
• Theorem 7
• Definition 3: Eclipsing with perfect CSI