Game Theoretic Strategies Design for Monostatic Radar and Jammer Based on Mutual Information

Kang Li, Bo Jiu, Hongwei LiuIEEE Access 2019 · DOI

Abstract

This paper applies game theory to model the high-stakes conflict between "smart" monostatic radars and intelligent jammers. By using Mutual Information as a utility metric, the authors derive optimal strategies for waveform design and jamming power allocation, proving that even when a stable Nash Equilibrium doesn't exist, hierarchical "Stackelberg" strategies provide a reliable performance floor for electronic warfare.

Introduction: The Intelligent Electronic Battlefield

In traditional electronic warfare, Electronic Counter-Countermeasures (ECCM) for radars and Electronic Countermeasures (ECM) for jammers were designed against "dumb" opponents—systems with fixed, predictable behaviors. However, the rise of cognitive radar and smart jammers has turned this into a dynamic, intelligent conflict.

The challenge is that when both sides are smart, they adapt to each other. If a radar changes its frequency, a smart jammer senses this and follows. If the jammer concentrates power in a specific band, the radar might shift its energy elsewhere. This research addresses the need for a mathematical framework that accounts for this mutual intelligence, specifically focusing on how to design the radar's transmit waveform and the jammer's power spectrum density (PSD) to maximize or minimize the information gained about a target.

System Model and Game Formulation

The authors frame the conflict as a Two-Person Zero-Sum (TPZS) game. They use Mutual Information (MI) between the received signal and the target's impulse response as the utility function. The radar wants to maximize MI to better "see" the target, while the jammer wants to minimize it.

Figure 1: Signal model of the monostatic radar and jammer system.

As shown in the signal model above, the system considers a monostatic radar transmitting a signal $x(t)$ through a channel with impulse response $h(t)$. The received signal $z(t)$ is corrupted by both additive white Gaussian noise (AWGN) $n(t)$ and a jamming signal $j(t)$. The combined signal passes through an ideal bandpass filter (BPF) $b(t)$ to produce the final output $y(t)$.

Two specific game structures are analyzed:

  1. Stackelberg Game (Hierarchical): One player is the "leader" (their strategy is intercepted first) and the other is the "follower." The authors solve for the Stackelberg Equilibrium (SE) using a two-stage optimization process involving Lagrange multipliers.
  2. Egalitarian Game (Symmetric): Both players choose strategies independently without knowing the other's move in advance. The authors investigate the conditions required for a Nash Equilibrium (NE)—a state where neither player can benefit by changing their strategy alone.

Key Insights and Ideas

  • The "Water-Filling" Strategy: Both the radar and the jammer naturally adopt "water-filling" strategies. They "pour" their available energy into the frequency bands where the target's Energy Spectral Variance (ESV) is highest.
  • The NE Existence Condition: A crucial finding is that a Nash Equilibrium only exists in the egalitarian game if the target's ESV is constant across the bandwidth. If the target's signature varies by frequency (which is true for most real-world complex targets), a pure-strategy NE does not exist.
  • SE as a "Safe" Strategy: When an NE doesn't exist, the authors argue that Stackelberg strategies are the next best thing. For the radar, the SE provides a "gain-floor" (a guaranteed minimum performance), and for the jammer, it provides a "loss-ceiling" (a limit on how much information the radar can gather).
  • Leader vs. Follower: The paper demonstrates that the strategies differ slightly depending on who leads. If the jammer leads, its optimal move is often to spread power more uniformly to cover its bases, whereas a leading radar must be more calculated about its energy distribution.

Simulation Results

Simulation results show that as radar power increases, the waveform spectrum becomes wider to capture more information. Conversely, as jamming power increases, the radar is forced to concentrate its energy into narrower, high-ESV bands to "punch through" the noise.

The authors numerically proved the "gain-floor" and "loss-ceiling" concepts. By comparing SE strategies against uniform and random strategies, they showed that SE consistently protects the player from the "worst-case" behavior of an intelligent opponent. The derived strategies successfully optimized Mutual Information compared to non-game-theoretic approaches, validating the two-stage optimization method.

Conclusion and Future Work

While this research provides a robust theoretical foundation, calculating optimal strategies in real-time during a split-second electronic warfare engagement remains a hurdle. Furthermore, the paper focuses on monostatic radar; real-world scenarios often involve bistatic or multistatic configurations (MIMO), which would significantly complicate the game dynamics. Integrating complex, non-Gaussian clutter environments is a suggested next step for future research.

Summary updated

BibTeX
@article{li2019game,
  author = {Kang Li and Bo Jiu and Hongwei Liu},
  title = {{Game Theoretic Strategies Design for Monostatic Radar and Jammer Based on Mutual Information}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2920398},
}
Figure