Bayesian Beamforming for Integrated Sensing and Communication Systems

Zongyao Zhao, Zhenyu Liu, Wei Dai, Xinke Tang, Xiao–Ping Zhang, Yuhan DongarXiv 2025 · DOI · arXiv

TL;DR

This paper tackles the challenge of designing beamforming for integrated sensing and communication (ISAC) systems when the target's properties are uncertain. They introduce a new metric, Expected Detection Probability (EPd), and a Bayesian beamforming approach to optimize it, outperforming existing methods in robust target detection.

What Problem Does This Paper Address?

Integrated Sensing and Communication (ISAC) systems aim to combine radar sensing and wireless communication capabilities within a single platform, promising enhanced efficiency and new applications. A crucial component of ISAC is beamforming, which directs the transmitted signal to optimize both communication and sensing performance. However, a significant challenge arises from the inherent uncertainty in sensing targets. In real-world scenarios, the target's scattering coefficient and its precise location (represented by azimuth angle) are often unknown and can vary. Existing beamforming designs typically assume these target parameters are known, which limits their effectiveness in practical, dynamic environments. This paper addresses this gap by developing a beamforming strategy that explicitly accounts for target uncertainty, aiming to maximize the average detection performance over the possible target states.

Methods

The core of this paper lies in its novel approach to handling target uncertainty.

  • Probabilistic Target Modeling: Instead of assuming fixed target parameters, the authors model the target's scattering coefficient and azimuth angle as random variables with known probability distributions.
  • Expected Detection Probability (EPd): A new metric, EPd, is introduced to quantify the average sensing performance from a Bayesian perspective. This metric integrates the probability of detection over the uncertainty of the target's parameters.
  • Bayesian Beamforming Formulation: The problem is framed as maximizing this EPd metric, subject to constraints on the communication system's performance (e.g., Signal-to-Interference-plus-Noise Ratio, SINR) and the base station's power budget.
  • Successive Convex Approximation and Semidefinite Relaxation (SCA-SDR): The resulting optimization problem is non-convex and complex. To solve it, the authors employ a combination of SCA and SDR techniques. SCA iteratively approximates the non-convex objective function with a convex one, while SDR relaxes certain constraints (like rank-1 constraints) to make the problem tractable. An iterative algorithm (Algorithm 1) is proposed to find the optimal beamformer.

Datasets

Not discussed in the paper. The evaluation is based on simulated scenarios with specified parameters for communication users, target properties, and noise levels.

Key Insights and Ideas

  • Uncertainty is Key: Explicitly modeling target uncertainty is crucial for robust ISAC performance, especially in practical scenarios. Assuming known parameters leads to suboptimal detection.
  • EPd as a Better Metric: The Expected Detection Probability (EPd) provides a more realistic and informative measure of sensing performance in uncertain environments compared to traditional metrics.
  • Bayesian Beamforming Pays Off: A Bayesian approach that optimizes EPd leads to beamforming strategies that are more resilient to target variations and achieve better detection rates.
  • SCA-SDR for Tractability: The development of an SCA-SDR algorithm is a significant methodological contribution, enabling the solution of a complex, non-convex optimization problem in this domain.

Key Results

  • The proposed Bayesian beamforming scheme, optimized for EPd, consistently outperforms baseline schemes (e.g., maximizing omnidirectional coverage, maximizing SINR) in terms of expected detection probability.
  • The scheme demonstrates robust detection performance even when target parameters are unknown and random.
  • The proposed method can achieve a desired level of EPd with significantly less transmit power compared to other schemes, especially as transmit power increases.
  • The SCA-SDR algorithm effectively improves the EPd through iterations and exhibits good convergence performance.

Limitations and Future Work

  • Discretization Approximation: The paper mentions approximating the integral in the EPd calculation by discretizing the target distribution. While effective, this introduces an approximation that could be refined.
  • Known Distributions: The current approach assumes the probability distributions of the target's parameters are known. Future work could explore scenarios where these distributions are also uncertain or need to be learned.
  • Complexity: While the SCA-SDR algorithm provides a solution, the computational complexity is still significant, especially for large-scale systems. Further optimization of the algorithm might be beneficial.
  • System Complexity: The paper focuses on a specific ISAC scenario. Extending the framework to more complex scenarios, such as multiple targets with correlated uncertainties or more sophisticated communication models, would be valuable.

Summary updated

Abstract

The uncertainty of the sensing target brings great challenge to the beamforming design of the integrated sensing and communication (ISAC) system. To address this issue, we model the scattering coefficient and azimuth angle of the target as random variables and introduce a novel metric, expected detection probability (EPd), to quantify the average detection performance from a Bayesian perspective. Furthermore, we design a Bayesian beamforming scheme to optimize the expected detection probability under the limited power budget and communication performance constraints. A successive convex approximation and semidefinite relaxation-based (SCA-SDR) algorithm is developed for the complicated non-convex optimization problem corresponding to the beamforming scheme. Simulation results show that the proposed scheme outperforms other benchmarks and exhibits robust detection performance when parameters of the target are unknown and random.
BibTeX
@misc{zhao2025bayesian,
  author = {Zongyao Zhao and Zhenyu Liu and Wei Dai and Xinke Tang and Xiao–Ping Zhang and Yuhan Dong},
  title = {{Bayesian Beamforming for Integrated Sensing and Communication Systems}},
  year = {2025},
  eprint = {2502.07012},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2502.07012},
}
Figure