Key Features

What you can do

Multimodal Grasping Mechanism

Combines layer jamming and tendon-driven mechanisms to adaptively grasp objects in cluttered environments.

Deep Multistage Learning

Implements a deep learning scheme using PyTorch to improve grasping performance in simulation.

CoppeliaSim Integration

Supports simulation scenes with Vortex physics engine for both lightly and highly cluttered scenarios.

Python-Based Codebase

Written in Python 3 with dependencies including PyTorch, NumPy, OpenCV, SciPy, and Matplotlib for ease of extension.

Dataset Support

Includes datasets for training grasping models in cluttered environments, requiring manual download and setup.