摘要：由于传感器、模型架构和训练数据各不相同，协同感知（CP）智能体之间存在的域差异，已成为新兴 CP 框架面临的主要挑战。以往工作通过模型再训练或针对每类智能体设置解释器，将特征对齐到统一空间，以缓解这一挑战。这些策略：（a）需要访问邻居智能体的配置；（b）未能完全解决实时 CP 部署问题；（c）对于运行时加入的未知智能体，泛化能力较差。为应对这些挑战，我们提出 PEARL，即面向匿名、实时、轻量级异构协同感知的提示嵌入（Prompt-Embedding）框架。PEARL 支持多个 CP 解释器，并利用两个并行训练的轻量级多尺度解释器，在运行时为新加入的智能体选择其中之一：稀疏检测（LWSD）解释器，用于对齐显著区域以支持协同检测；稠密、域不变（LWDDI）解释器，用于生成与智能体无关的特征，以实现快速解释器选择。两个解释器均使用低秩视觉提示，以降低计算量、存储需求和模型复杂度。在模拟数据集（OPV2V、V2XSet）和真实数据集（DAIR-V2X）上的大量实验表明，PEARL 可泛化至模拟和真实世界的协同驾驶场景。其实时模型选择策略相较随机选择基线带来 8.2% 的平均精度（AP）提升，平均运行时间为 1.67 ms。尽管 PEARL 主要面向实时 CP 设计，但在传统离线训练条件下，其平均 AP 仍比最先进的异构 CP 框架高出 5.6%，同时可将通信成本最多降低 34.7 倍。同样重要的是，PEARL 无需共享智能体的配置或模型设置，因此能够保护可能属于专有或私密的信息。这些结果表明，PEARL 是一套可扩展且实用的异构协同感知框架。

Abstract:Heterogeneity across Collaborative Perception (CP) agents is a major challenge for emerging CP frameworks due to domain gaps from differing sensors, architectures, and training data. Prior works mitigate this challenge by aligning features in a unified space via model retraining or per-agent-type interpreters. These strategies (a) require access to neighbor configurations, (b) do not fully address real-time CP deployment, and (c) generalize poorly to unseen agents joining at run time. To overcome these challenges, we present PEARL, a Prompt-Embedding framework for Anonymous and Real-time Lightweight heterogeneous CP. PEARL supports multiple CP interpreters and selects one for a new-joining agent in real time using two lightweight, multi-scale interpreters trained in parallel: a sparse-detection (LWSD) interpreter that aligns salient regions for cooperative detection, and a dense, domain-invariant (LWDDI) interpreter that produces agent-invariant features for fast interpreter selection. Both interpreters use low-rank visual prompts to reduce computation, storage, and model complexity. Extensive experiments on simulated (OPV2V, V2XSet) and real (DAIR-V2X) datasets show that PEARL generalizes across simulated and real-world cooperative driving scenarios. Its real-time model-selection strategy yields an 8.2% Average Precision (AP) gain over a random-selection baseline while running in 1.67 ms on average. Although primarily designed for real-time CP, PEARL also outperforms state-of-the-art heterogeneous CP frameworks under traditional offline training by 5.6% AP on average while reducing communication cost by up to 34.7 times. Equally important, PEARL does not require sharing agents' configurations or model settings, thereby protecting information that may be proprietary or private. These results establish PEARL as a scalable and practical framework for heterogeneous collaborative perception.