Development of robotic grasping of object detection model

Robotic grasping stands as a foundational task within the realm of robotics, facilitating effective interaction and manipulation of objects in their environment. In recent years, the integration of machine learning has revolutionized robotic grasping, empowering robots to grasp directly from r...

Full description

Saved in:
Bibliographic Details
Main Author: Koh, Aloysius Jun Jie
Other Authors: Cheah Chien Chern
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2024
Subjects:
Online Access:https://hdl.handle.net/10356/177230
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Robotic grasping stands as a foundational task within the realm of robotics, facilitating effective interaction and manipulation of objects in their environment. In recent years, the integration of machine learning has revolutionized robotic grasping, empowering robots to grasp directly from raw sensory data. This paper aims to explore a machine learning-based approach to robotic grasping, with a specific focus on leveraging basic shape features to discern potential grasping points. It provides a comprehensive documentation of the process involved in training the YOLOv5 robotic grasping model, starting from the stage of image collection, proceeding to the development of algorithms aimed at identifying feasible grasping points based on the detected data, and culminating in the implementation of the model and algorithms onto an actual robotic arm for simulation.