Skip to content

论文arXiv 2611.234993 天前

BiCAT双臂协调动作Token化:仿真成功率提升17.6%且推理仅40ms ​

本文由 AI 根据公开内容整理,可能有误,请以原文为准:BiCAT: Bimanual Coordination Action Tokenization(arXiv 2611.23499)

核心内容 ​

现有自回归视觉-语言-动作(VLA)模型在处理双臂操作时,常将双臂动作拼接为单一向量,导致共性运动模式无法共享且跨臂依赖纠缠。为此,研究提出BiCAT(Bimanual Coordination Action Tokenization),旨在通过改进动作Tokenizer接口来显式建模双臂协调机制。该方法在RoboTwin 2.0的10项仿真任务及3种本体的6项真机任务中进行了全面验证,结果显示其在所有对比Tokenizer中取得最高成功率,尤其在需紧密配合的任务中优势显著。

技术上,BiCAT采用共享运动表示并结合对称交叉注意力机制与联合解码器,使每臂Token生成均条件依赖于对侧手臂。策略网络并行预测各臂Token,遵循“先生成左臂、右臂以此为条件”的顺序。实验表明,相比8个基线中的最强者,BiCAT在仿真中平均成功率提升17.6个百分点,同时保持极低计算开销,每个动作块(action chunk)推理仅需40.5ms,证明了在Token化阶段引入结构化归纳偏置对提升双臂具身智能性能的关键作用。

关键看点 ​

  • 架构创新:采用共享运动表示与对称交叉注意力机制,解耦双臂共性模式与跨臂依赖,替代传统向量拼接法
  • 性能跃升:在RoboTwin 2.0仿真测试中,平均成功率较8个基线中最强者提升17.6个百分点,协作任务增益最大
  • 高效推理:优化Token预测流程,支持双臂Token并行生成,单个动作块推理延迟低至40.5ms,满足实时控制需求
  • 泛化验证:覆盖10项仿真任务及3种不同本体的6项真机任务,证实该Token化方案在异构硬件上的通用性

工程与产业观察 ​

该工作确立了动作Tokenizer作为双臂协调建模核心组件的工程价值。对于具身算法工程师,BiCAT证明了无需增加骨干网络参数量,仅通过在离散化阶段引入对称注意力与共享表征等结构化先验,即可大幅降低VLA模型学习双臂耦合策略的难度。其40.5ms的低延迟特性使其具备端侧部署潜力,为高精度双臂灵巧操作提供了高性价比的基准方案。

英文摘要 (Abstract) ​

Autoregressive vision-language-action (VLA) models predict discrete action tokens with the next-token objective of their vision-language backbone, so the action tokenizer defines the interface between the policy and the robot. In bimanual manipulation, the two arms share a repertoire of movements yet coordinate through complementary roles, and an action representation should capture both. However, most existing tokenizers encode both arms as a single concatenated action vector, in which motion patterns common to the two arms are not shared and cross-arm dependencies remain entangled within each token. The policy must therefore learn both aspects implicitly. We propose Bimanual Coordination Action Tokenization (BiCAT), which shares one motion representation across arms while conditioning each arm's tokens on its partner through symmetric cross-arm attention and a joint decoder. The policy then predicts all tokens of each arm in parallel, generating the left arm first and the right arm conditioned on it. Across ten RoboTwin 2.0 tasks and six real-world tasks on three embodiments, BiCAT achieves the highest success among the compared tokenizers. In simulation, it improves mean success by 17.6 percentage points over the strongest of eight baselines, with the largest gains on tasks that require close coordination between the arms, while requiring only 40.5 ms per action chunk. These results suggest that the action tokenizer is a natural place to model bimanual coordination.

作者:Xuanwei Liu, Guipu Cao, Zhongyi Luo, Jiafei Duan

基于本页技术方向关联的整机参数、核心零部件、实测报告与同赛道选型横评。

外部来源

本页事实对应的公开页面,以原文为准。(点击将离开本站)

  1. arXiv 2611.23499:BiCAT: Bimanual Coordination Action Tokenization ↗访问 2026-11-19