SlipSense 多模态触觉学习实现低延迟通用滑移检测
SlipSense: Multimodal Tactile Learning for Low-Latency and Generalized Slip Detection
这个新框架 Slipsense 很实用，用两种传感器数据做滑移检测，比单独用一种更准，延迟还低，还能跨不同机器手和物体用，挺有意思的。
SlipSense 是一个基于 TacV5 传感器的多模态滑移检测框架，通过 32x32 压阻阵列和 3 轴 MEMS 加速度计捕捉压力分布和摩擦振动，在 240Hz 频率下进行模态编码、传感器内融合和跨模态注意力计算。在包含 1.4 百万帧、37 个物体的数据集上，它实现了 96.7% 的 Macro F1 分数，在 23.1ms 内检测到 76% 的滑移事件，且仅用 UMI 数据训练就能零样本迁移到 Tesollo 机器手。
SlipSense: Multimodal Tactile Learning for Low-Latency and Generalized Slip Detection
Slip detection is fundamental to dexterous manipulation, yet existing systems often lack precise characterization of detection latency and cross-platform generalization. We present SlipSense, a multimodal tactile slip-detection framework built on TacV5, a compact sensor integrating a $32 \times 32$ piezoresistive array operating at 240 Hz and a 3-axis MEMS accelerometer operating at 8 kHz. The piezoresistive array captures spatial pressure distributions, while the accelerometer captures friction-induced vibrations, providing complementary slip cues. The framework performs modality-specific encoding, intra-sensor fusion, and cross-modal attention with causal temporal prediction at 240 Hz. Experiments on a dataset of 1.4 million frames spanning 37 objects demonstrate the complementarity of the two modalities. SlipSense achieves 96.7% Macro F1 with a false-positive rate below 1.6%, detecting 76% of slip events within 23.1 ms. When trained solely on UMI data, SlipSense generalizes zero-shot to a Tesollo dexterous hand, transferring across unseen objects, distinct sensor units, and robotic platforms without retraining.
