Joint Node Selection and Resource Allocation Optimization for Cooperative Sensing with a Shared Wireless Backhaul

Mingxin Chen, Ming-Min Zhao, An Liu, Min Li, Qingjiang Shi

Abstract

Future 6G networks aim to integrate sensing and communication, enabling base stations to act as a distributed radar system. However, sharing high-resolution sensing data across wireless backhaul networks creates significant overhead. This paper proposes a Hybrid Information-Signal Domain Cooperative Sensing (HISDCS) framework. By jointly optimizing node selection and bit allocation using Karhunen-Loève Transformation (KLT) and a successive convex approximation algorithm, the framework achieves high-accuracy localization while reducing communication overhead by over 30% compared to traditional methods.

Introduction

In the evolution toward 6G, cooperative sensing allows multiple base stations (BSs) to collaborate in tracking targets like autonomous vehicles or drones. Traditionally, this is done via two extremes: Information-Domain Cooperative Sensing (IDCS), which is bandwidth-efficient but less accurate, and Signal-Domain Cooperative Sensing (SDCS), which is highly accurate but demands massive backhaul capacity. This paper addresses the challenge of finding a "Goldilocks" zone—achieving SDCS-level precision without overwhelming the network.

System Model

The system consists of a transmitter, a target, and $N$ receivers connected to a Fusion Center (FC) via a limited-capacity wireless backhaul. The transmitter sends a probing signal, which reflects off the target. The receivers capture these echoes and must decide what information to send to the FC to estimate the target's location $\theta$.

The system model for cooperative sensing with a limited backhaul network.

The HISDCS Framework

The authors introduce HISDCS (Hybrid Information-Signal Domain Cooperative Sensing). This framework operates in four distinct steps to balance accuracy and efficiency:

  1. Local Estimation (Information Domain): Each receiver performs a Maximum Likelihood (ML) estimation to find a rough time-of-arrival (TOA) and signal amplitude.
  2. Resource Optimization: The system determines which nodes should participate and how many bits should be allocated to each signal snippet to meet a specific accuracy requirement (defined by the Cramér-Rao Lower Bound).
  3. KLT Quantization (Signal Domain): Instead of sending the full raw signal, receivers extract a small "snippet" around the estimated TOA. This snippet is compressed using the Karhunen-Loève Transformation (KLT), which is mathematically optimal for preserving the most critical features of Gaussian signals.
  4. Cooperative Sensing: The FC receives both the rough estimates and the quantized signal snippets to produce a high-precision final location estimate $\hat{\theta}$.
The four-step workflow of the proposed HISDCS framework.

Technical Innovations

  • Joint Optimization: The paper demonstrates that optimizing bit-rate and node selection in isolation is sub-optimal. The authors developed a Matrix-inequality Constrained Successive Convex Approximation (MCSCA) algorithm to solve the joint problem efficiently.
  • Greedy Node Selection: To further save bandwidth, the system employs a greedy strategy to prune redundant or unreliable nodes that do not significantly contribute to the localization accuracy.

Evaluation and Key Results

The framework was evaluated using numerical simulations of a standard 6G environment (3.55 GHz Gaussian pulse, linear topology of 5 receivers).

  • Efficiency Gains: HISDCS reduced communication overhead by 30.1% at 0 dB SNR compared to schemes that do not utilize node selection.
  • Accuracy: The performance of the hybrid design consistently outperformed traditional TOA and TOA-RSS (Received Signal Strength) baselines. With minimal bit allocation, the system's accuracy approached that of "Ideal SDCS," which assumes an infinite backhaul capacity.
  • The Power of Snippets: By sending only the signal samples around the estimated time delay, the system captures the essential data needed for fine-tuning without the overhead of the full waveform.

Conclusion and Future Work

This research proves that smart compression and joint resource management can bridge the gap between efficiency and precision in 6G sensing. However, several challenges remain for real-world deployment. Future work must address multipath interference in urban environments, which can distort signal snippets, and the requirement for extremely tight synchronization between base stations. Additionally, expanding the framework to handle multi-target tracking and integrating MIMO antenna arrays are key areas for further exploration.

Summary updated

BibTeX
@article{chen2025joint,
  author = {Mingxin Chen and Ming-Min Zhao and An Liu and Min Li and Qingjiang Shi},
  title = {{Joint Node Selection and Resource Allocation Optimization for Cooperative Sensing with a Shared Wireless Backhaul}},
  journal = {IEEE Trans. Signal Process.},
  year = {2025},
  doi = {10.1109/tsp.2024.3516709},
}
Figure