BEHAVIOR Robot Suite (BRS) is a framework developed by Fei-Fei Li's team at Stanford University for learning whole-body manipulation to complete daily household tasks.
BEHAVIOR Robot Suite
Introduction
BEHAVIOR Robot Suite (BRS) is a framework developed by Fei-Fei Li's team at Stanford University to enable robots to perform complex household tasks through whole-body manipulation. It addresses key challenges in robotics, such as bimanual coordination, stable navigation, and extensive end-effector reachability, making it highly effective in real-world environments.
Key Features
- Whole-body manipulation: Completes tasks like carrying heavy objects, opening doors, and cleaning.
- Efficient data collection: Uses JoyLo, a low-cost teleoperation interface, for high-quality data collection.
- Advanced learning algorithm: WB-VIMA models hierarchical whole-body movements and uses multimodal data for precise control.
- Real-world adaptability: Interacts with everyday objects in complex environments.
- Fault recovery: Automatically detects and corrects errors during operations.
Technical Details
- JoyLo: A low-cost teleoperation interface using hardware like Nintendo Joy-Con for precise whole-body control.
- WB-VIMA: An imitation learning algorithm that decomposes whole-body movements and uses self-attention to aggregate multimodal data.
Use Cases
- Household chore automation: Cleaning rooms, placing items on shelves, and putting away dishes.
- Garbage disposal: Moving trash bags from indoors to outdoor bins.
- Clothing organization: Taking clothes out of the wardrobe and placing them on the sofa.
- Bathroom cleaning: Cleaning the toilet and pressing the flush button.
- Daily item management: Opening the dishwasher and managing items.
Getting Started
Visit the official website for documentation, code, and datasets.