Robbyant, an embodied AI company within Ant Group, released the open-source tactile model LingBot-Touch on November 10.

The model helps robots process touch signals from different sensor types through a shared representation.

The company said LingBot-Touch was pre-trained on nearly 5,000 hours of real-world tactile data and supports magnetic, visual, and piezoresistive sensing.

In 26 tests using two robot hands and four action models, average task success rates improved by about 14 to 26 percentage points compared with setups without tactile input, according to Robbyant.

Robbyant said it aims to build “one brain for all robots.” LingBot-Touch is meant to make tactile systems work across different sensor designs.

The release is part of Ant Group’s broader open-source robotics push to move robots from lab demos into real-world work.

Other open tactile efforts include tools to benchmark sensors and projects that offer scalable tactile systems.

🔗 Source: Robbyant

Recent Robbyant developments