Timeliness-Oriented Scheduling and Resource Allocation in Multi-Region Collaborative Perception

TL;DR

This work addresses timeliness and resource constraints in multi-region collaborative perception (CP) by formulating a timeliness-aware scheduling problem and solving it with a Lyapunov-based policy. The authors introduce the Timeliness-Aware Multi-region Prioritized (TAMP) scheduling algorithm, which uses an empirical penalty to jointly account for AoI and communication volume, and derives a per-slot priority index to guide region selection and adaptive data transmission. By modeling the penalty via a tractable surrogate and leveraging virtual queues, TAMP achieves online, resource-aware decisions that balance perception accuracy with communication and computation costs; experiments on the Roadside Cooperative (RCooper) dataset show AP gains up to 27% over strong baselines in intersection and corridor scenarios. The empirical penalty fitting and the Split-based adaptive transmission design enable practical, scalable CP in dynamic environments, signaling meaningful improvements for autonomous driving and large-scale smart-city perception systems.

Abstract

Collaborative perception (CP) is a critical technology in applications like autonomous driving and smart cities. It involves the sharing and fusion of information among sensors to overcome the limitations of individual perception, such as blind spots and range limitations. However, CP faces two primary challenges. First, due to the dynamic nature of the environment, the timeliness of the transmitted information is critical to perception performance. Second, with limited computational power at the sensors and constrained wireless bandwidth, the communication volume must be carefully designed to ensure feature representations are both effective and sufficient. This work studies the dynamic scheduling problem in a multi-region CP scenario, and presents a Timeliness-Aware Multi-region Prioritized (TAMP) scheduling algorithm to trade-off perception accuracy and communication resource usage. Timeliness reflects the utility of information that decays as time elapses, which is manifested by the perception performance in CP tasks. We propose an empirical penalty function that maps the joint impact of Age of Information (AoI) and communication volume to perception performance. Aiming to minimize this timeliness-oriented penalty in the long-term, and recognizing that scheduling decisions have a cumulative effect on subsequent system states, we propose the TAMP scheduling algorithm. TAMP is a Lyapunov-based optimization policy that decomposes the long-term average objective into a per-slot prioritization problem, balancing the scheduling worth against resource cost. We validate our algorithm in both intersection and corridor scenarios with the real-world Roadside Cooperative perception (RCooper) dataset. Extensive simulations demonstrate that TAMP outperforms the best-performing baseline, achieving an Average Precision (AP) improvement of up to 27% across various configurations.

Figures (11)

• Figure 1: An illustration of the system architecture, where a BS manages multiple CP regions.
• Figure 2: Timeliness-oriented penalty function with AoI and the communication volume in a region. The AoI $h_k$ increases in a staircase manner over time and is reset only upon the completion of a CP task. The $m$-th schedule begins at slot $k_m$ with the allocated communication volume (Comm) $\bm{b}_{k_m}$, and the task is completed upon slot $k'_m$. The $m$-th interval is defined as the duration between the completion time of two tasks, from $k'_m$ to $k'_{m+1}$. Note that $h_{k_m}$ indicates the AoI at the $m$-th scheduling time.
• Figure 3: The workflow of the multi-region CP system.
• Figure 4: Two typical roadside CP scenarios of RCooper datasethao2024rcooper.
• Figure 5: Performance fitting in the intersection scenario, which depicts the joint impact of AoI and communication volume (in $\log_2(\text{Bytes})$) on AP@0.5.

Theorems & Definitions (9)

• Lemma 1
• proof
• Remark 1
• Lemma 2
• proof
• Theorem 1
• proof
• Remark 2
• Remark 3