Robotic Grasping

Research Intern - Lab of Autonomous Robotics and Systems, Fudan University

:mag: ROS(Robot Operating System), C++, Python

  • Built software on an Aubo-i5 robot arm with Robotiq gripper(USB), Chishine3D RGB-D camera and a UR5 robot arm with custom gripper(Modbus), Intel RealSense D435 RGB-D camera to grasp various fruits like apples, oranges, and bananas placing on the table using the open-source framework, ROS Melodic on Ubuntu 18.04. Implemented object localization and classification using YOLOv4(You Only Look Once: Unified, Real-Time Object Detection). Experienced hand-eye calibration and implemented grasping procedure including motion planning using MoveIt.
  • Developed low-cost solutions like edge-based geometric shape detection and template matching to perform object localization, classification and grasp pose estimation on 2D RGB images.
  • Explored open-source Cartographer SLAM simulation.