通过瓶颈潜在重构实现无需触觉仿真的触觉Sim2Real
Tactile Sim2Real without Tactile Simulation via Bottlenecked Latent Reconstruction
- University of Michigan（密歇根大学）
- Google DeepMind（谷歌DeepMind）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
该研究提出SBLR框架，无需触觉仿真，通过两阶段策略训练和未配对数据对齐，在插销插入等任务上实现高零样本成功率，性能优于物理仿真基线。
AI中文摘要：
机器人传感器设计，尤其是触觉传感器，种类繁多且发展迅速。在仿真中对每种传感器进行建模需要大量领域专业知识，而计算近似会降低仿真信号的保真度。我们提出Sim2Real via Bottlenecked Latent Reconstruction（SBLR）框架，该框架完全避免了特定传感器的仿真，具体方法为：（1）在易于构建的仿真原生 oracle 传感器上训练策略，无需对任何特定传感器建模（例如，我们使用点云和指尖力作为触觉 oracle）；（2）在推理时将真实传感器的潜在嵌入与 oracle 传感器的潜在嵌入对齐。策略训练分为两个阶段：策略首先从 oracle 传感器潜在信息中学习，然后通过瓶颈潜在重构使其适应使用真实传感器而非 oracle 时预期的信息损失。oracle 传感器与真实传感器之间的对齐通过在仿真和真实世界中收集的未配对随机游戏数据学习，使用基于校正流的变换网络，该网络在最近邻伪对上进行训练。在三个接触丰富任务上的仿真实验表明，SBLR 的性能与直接访问触觉仿真的 oracle 相当或接近。在使用GelSight Mini和DIGIT传感器进行的插销插入和齿轮啮合的硬件实验中，SBLR实现了85%-97.5%的零样本成功率，无需任何特定传感器建模或校准，比基于物理的触觉仿真基线性能高出7.5%-15%。
英文摘要：
Robot sensor designs, particularly tactile sensors, are highly diverse and evolve rapidly. Modeling each sensor in simulation demands substantial domain expertise and computational approximations can degrade the fidelity of the simulated signals. We propose Sim2Real via Bottlenecked Latent Reconstruction (SBLR), a framework that avoids sensor-specific simulation entirely by (1) training policies on a simulator-native oracle sensor that is easy to construct without modeling any particular sensor (e.g. we use a point-cloud and finger-tip forces as a tactile oracle), and (2) aligning real sensor latent embeddings to those of the oracle sensor at inference time. Policy training proceeds in two-stage: the policy first learns from the oracle sensor latents, then a bottlenecked latent reconstruction adapts it to the information loss expected when using the real sensor instead of the oracle. The alignment between oracle and real sensor is learned from unpaired random-play data collected in both simulation and the real world, using rectified-flow-based transformation networks trained on nearest-neighbor pseudo-pairs. Simulation experiments on three contact-rich tasks show that SBLR matches or approaches the performance of an oracle with direct access to tactile simulation. Hardware experiments on Peg Insertion and Gear Meshing with GelSight Mini and DIGIT sensors demonstrate 85-97.5% zero-shot success without requiring any sensor-specific modeling or calibration, outperforming a physics-based tactile simulation baseline by 7.5-15%.
