Skip to content
Hot eventLive

RATE:面向接触密集型机器人操作的风险感知触觉编码

1 reports1 sources16 hr ago updated

Get the story

AI overview

研究者提出 Risk-Aware Tactile Encoding(RATE),用于学习编码任务条件交互风险的触觉表示。该方法通过历史条件预测捕捉交互上下文,并以警报监督将上下文与任务条件风险关联,再经轻量残差适配器补充传统触觉特征。仿真与真实世界实验显示任务成功率显著提升,消融实验验证了警报引导学习与预测时序建模的互补收益。目前报道仅涉及方法框架与实验结果,未披露具体数据集、量化指标或后续部署信息。

Generated from reports · updated 16 hr ago

Timeline

Follow the coverage from different angles.

Oct 5, 2026
  1. arXiv · Robotics
    RATE:面向接触密集型机器人操作的触觉风险感知编码

    研究者提出 Risk-Aware Tactile Encoding(RATE),用于学习编码任务条件交互风险的触觉表示。该方法通过历史条件预测捕捉交互上下文,并以警报监督将上下文与任务条件风险关联,再经轻量残差适配器补充传统触觉特征。仿真与真实世界实验显示任务成功率显著提升,消融实验验证了警报引导学习与预测时序建模的互补收益。

Heat trend

Current heat 7·Comparable peak 10(Oct 5)·Comparable change over 24 hours –

02.557.510Oct5Oct5Oct5Oct6

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.