DexSkills: Skill Segmentation Using Haptic Data for Learning Autonomous Long-Horizon Robotic Manipulation Tasks

Xiaofeng Mao,Gabriele Giudici,Claudio Coppola,Kaspar Althoefer,Ildar Farkhatdinov,Zhibin Li,Lorenzo Jamone

2024-05-06

Abstract:Effective execution of long-horizon tasks with dexterous robotic hands remains a significant challenge in real-world problems. While learning from human demonstrations have shown encouraging results, they require extensive data collection for training. Hence, decomposing long-horizon tasks into reusable primitive skills is a more efficient approach. To achieve so, we developed DexSkills, a novel supervised learning framework that addresses long-horizon dexterous manipulation tasks using primitive skills. DexSkills is trained to recognize and replicate a select set of skills using human demonstration data, which can then segment a demonstrated long-horizon dexterous manipulation task into a sequence of primitive skills to achieve one-shot execution by the robot directly. Significantly, DexSkills operates solely on proprioceptive and tactile data, i.e., haptic data. Our real-world robotic experiments show that DexSkills can accurately segment skills, thereby enabling autonomous robot execution of a diverse range of tasks.

Robotics

What problem does this paper attempt to address?

The paper attempts to address the problem of how to enable robots to efficiently perform long-horizon dexterous manipulation tasks in the real world. Specifically, the paper focuses on:

1. **Efficiency of Data Collection and Training**: Traditional learning methods require a large amount of demonstration data to train robots, which is not only time-consuming but also labor-intensive. Therefore, the paper proposes a new method that reduces the amount of data needed by decomposing long-horizon tasks into a series of primitive skills.

2. **Task Decomposition and Execution**: Long-horizon tasks often consist of multiple subtasks, which can be broken down into simpler, reusable skills. The paper proposes a framework that can identify and replicate these primitive skills from human demonstrations and decompose a long-horizon task into a series of primitive skills, allowing the robot to execute the entire task in one go.

3. **Utilization of Perceptual Data**: Traditional robotic operations typically rely on visual information, but in some cases, visual information may be unreliable or unavailable. The proposed method in the paper uses only proprioceptive and tactile data (i.e., force data), which is particularly important in contact-rich manipulation tasks and can provide valuable information about objects and the environment.

In summary, the main goal of the paper is to develop an efficient and reliable method that enables robots to autonomously perform complex long-horizon manipulation tasks by learning primitive skills, while reducing the dependence on a large amount of demonstration data.