Development and Testing of a Moblie Manipulation Robot

The graduation project for Bachelor of Science, from the class of 2024 of the Department of Electromechanical Engineering, University of Macau

by Cheney, Dechao Jiang and Bobby, Tong Zhang

Project Supervisor: PhD, FASME, Professor Qingsong Xu

Connect with us!

Cheney: [email protected] / [email protected]

Bobby: [email protected] / [email protected]

Introduction

Libot: Autonomous Mobile Manipulation Robot for Libraries

In this final year project, we proposed Libot. Libot is a mobile manipulation robot designed to automate tasks in library environments. Combining the capabilities of the UR5 robotic arm and MiR250 mobile platform, enhanced with a Realsense 435i depth camera, Libot autonomously navigates and interacts within library settings.

Feature

• Book Detection System: Utilizes the YOLO v8 deep learning model to recognize book spine labels from RGB images, facilitating efficient book cataloging.
• Deep-learning-based Object grasping: Integrated with the MoveIt platform through move group API to optimize book handling through precise pick-and-place operations. Deep-learning algorithm GPD by Andreas is used for grasp generation.
• SLAM Mapping and Navigation: Enhanced with Simultaneous Localization and Mapping (SLAM) to adapt to dynamic library layouts, improving navigation and operational efficiency. Mapping through hectoring mapping by TeamHector at TU Darmstadt.

For the Detailed Methodology, please click HERE.

Testing and Results

CLICK BELOW TO SEE DEMO VIDEO