Robot Navigation (Tutor: Zhanbo SHI, Wechat: 18366738320)

Introduction

The task of this project is to build a navigation system on a mobile robot.

Requirements

1> By entering an instruction (find a chair/table/door) via the keyboard, the robot can navigate to the corresponding object in the environment. If there are multiple instances of the same category, the robot can simply navigate to any one of them.

2> During the navigation process, the robot should be able to avoid static obstacles.

3> The navigation system should be deployed on a mobile robot in the real world.

Advanced Requirements

1> If the environment contains several instances of the same category, the robot must navigate to the specific instance requested by the instruction (find the red chair).

2> If there are no objects in the environment that match the instruction, the robot is able to give a feedback to the user.

Reference Materials

1. Object category list: https://github.com/niessner/Matterport/blob/master/metadata/mpcat40.tsv
2. Robot beginner tutorial: https://docs.taobotics.com/docs/handsfree/handsfree-en/Tutorial/Beginner/
3. Object detection: https://docs.ultralytics.com/models/yolov8/
4. Camera information: https://www.intelrealsense.com/depth-camera-d435i/
5. GPT for robot navigation: https://arxiv.org/abs/2306.17582

Created on: Nov. 07, 2024