Optimal Coordinated Transmit Beamforming for Networked Integrated Sensing and Communications

Gaoyuan Cheng, Yuan Fang, Jie Xu, Derrick Wing Kwan Ng

TL;DR

This paper provides a mathematical framework for optimizing "Networked ISAC"—a system where multiple base stations (BSs) coordinate their transmissions to simultaneously serve mobile users and detect targets (like a radar). By cleverly designing beamforming patterns for both data and dedicated sensing signals, the authors demonstrate how to maximize the probability of detecting a target in a specific area without compromising the quality of service for communication users.

What Problem Does This Paper Address?

As we move toward 6G, wireless networks are evolving from simple data pipes into "Integrated Sensing and Communications" (ISAC) systems. While previous research focused on single base stations (mono-static) or pairs (bi-static), this paper tackles the more complex Networked ISAC scenario.

The challenge is a classic trade-off: to detect a target, you want to blast energy toward it. However, in a networked environment, these "sensing beams" can act as interference for communication users (CUs). Furthermore, when multiple BSs are involved, they can potentially share "echo" signals (cross-links), but this requires precise time synchronization. The authors set out to find the optimal balance: how should multiple BSs shape their beams to maximize detection probability while ensuring every user maintains a minimum Signal-to-Interference-plus-Noise Ratio (SINR)?

Methods

The authors model a system with $L$ multi-antenna base stations and multiple single-antenna users. Their approach involves several key technical components:

  • Signal Design: BSs transmit a combination of information-bearing signals and dedicated "sensing-only" signals.
  • Receiver Modeling: They analyze two types of users: Type-I, who treat sensing signals as pure noise, and Type-II, who have the capability to cancel out the interference from known sensing signals.
  • Synchronization Scenarios:
    • Scenario I (Synchronized): BSs exploit both "direct links" (BS A $\rightarrow$ target $\rightarrow$ BS A) and "cross-links" (BS A $\rightarrow$ target $\rightarrow$ BS B).
  • Scenario II (Unsynchronized): BSs only utilize their own direct-link reflections.
  • Optimization Framework: The problem is formulated as a Max-Min Detection Probability problem. Because the resulting math is a non-convex Quadratically Constrained Quadratic Program (QCQP)—notoriously difficult to solve—the authors use Semi-Definite Relaxation (SDR). Crucially, they provide a rigorous proof that their SDR approach is "tight," meaning it yields the true optimal solution rather than just an approximation.

Datasets

This is a theoretical and simulation-based study; therefore, no external real-world datasets were used. Instead, the authors evaluated their algorithms using:

  • Standard Channel Models: Rayleigh fading for communication links and Line-of-Sight (LoS) models for sensing links.
  • Numerical Setup: A simulated environment with 3 BSs (each with 32 antennas) and varying numbers of users, with parameters like noise power and path loss exponents based on standard 3GPP-like settings.

Key Insights and Ideas

  • The Synchronization Multiplier: Time synchronization is a massive performance booster. By enabling the use of cross-links (Scenario I), the system gains "spatial diversity," significantly increasing the chances of detecting a target compared to the unsynchronized case.
  • Interference Cancellation is Essential: For Type-II receivers, dedicated sensing signals are "free" in terms of interference. This allows the BSs to be much more aggressive with sensing beams, leading to much higher detection probabilities when users and targets are in similar directions.
  • The "Sensing Signal" Paradox: For Type-I receivers (who can't cancel interference), the optimal strategy often involves sending zero dedicated sensing signals, relying instead entirely on the communication beams to do the "double duty" of sensing.
  • Max-Min Robustness: By maximizing the minimum detection probability over an area, the beamforming design ensures there are no "blind spots" in the targeted sensing zone.

Key Results

  • Superiority over Benchmarks: The proposed optimal beamforming consistently outperformed standard benchmarks, such as Zero-Forcing (ZF) beamforming, especially when communication requirements (SINR) were stringent.
  • Detection Gains: In synchronized scenarios, the detection probability was significantly higher (often reaching near 1.0 at moderate SINR thresholds) compared to unsynchronized setups, which struggled to maintain high detection as user requirements increased.
  • Angular Sensitivity: The simulations showed that when a user is located at the same angle as the target, the detection probability drops sharply for Type-I receivers but remains robust for Type-II receivers, proving the value of interference cancellation in ISAC.

Limitations and Future Work

  • Perfect CSI Assumption: The paper assumes the base stations have perfect Channel State Information (CSI) for the users. In reality, CSI is often imperfect, which could degrade beamforming performance.
  • Static Targets: The model focuses on detecting the presence of a target at a specific location. It does not account for moving targets (Doppler effects) or the tracking of multiple targets simultaneously.
  • Computational Complexity: While SDR is effective, solving large-scale QCQPs for networks with dozens of BSs and hundreds of users might become computationally prohibitive for real-time implementation.
  • Future Directions: The authors suggest extending this work to include robust beamforming under CSI uncertainty and exploring the coordination overhead required for Scenario I synchronization.

Summary updated

Abstract

This paper studies a multi-antenna networked integrated sensing and communications (ISAC) system, in which a set of multi-antenna base stations (BSs) employ the coordinated transmit beamforming to serve multiple single-antenna communication users (CUs) and perform joint target detection by exploiting the reflected signals simultaneously. To facilitate target sensing, the BSs transmit dedicated sensing signals combined with their information signals. Accordingly, we consider two types of CU receivers with and without the capability of canceling the interference from the dedicated sensing signals, respectively. In addition, we investigate two scenarios with and without time synchronization among the BSs. For the scenario with synchronization, the BSs can exploit the target-reflected signals over both the direct links (BS-to-target-to-originated BS links) and the cross-links (BS-to-target-to-other BSs links) for joint detection, while in the unsynchronized scenario, the BSs can only utilize the target-reflected signals over the direct links. For each scenario under different types of CU receivers, we optimize the coordinated transmit beamforming at the BSs to maximize the minimum detection probability over a particular targeted area, while guaranteeing the required minimum signal-to-interference-plus-noise ratio (SINR) constraints at the CUs. These SINR-constrained detection probability maximization problems are recast as non-convex quadratically constrained quadratic programs (QCQPs), which are then optimally solved via the semi-definite relaxation (SDR) technique.
BibTeX
@misc{cheng2023optimal,
  author = {Gaoyuan Cheng and Yuan Fang and Jie Xu and Derrick Wing Kwan Ng},
  title = {{Optimal Coordinated Transmit Beamforming for Networked Integrated Sensing and Communications}},
  year = {2023},
  eprint = {2307.05127},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2307.05127},
}
Figure