Skip to content

Indoor Navigation for a Differential-Drive Robot in an Unknown and Dynamic Environments

Notifications You must be signed in to change notification settings

rantengsky/Visual-Perception-based-Navigation

Repository files navigation

Visual-Perception-based-Navigation

Introduction

Authors: T. Ran, L. Yuan, J. Zhang

This project is a real-time indoor navigation system for the mobile robot. It takes the information of an monocular camera as input. The images captured by the camera will be classified by a CNN model. After the classification, the system converts the outputs to specific velocity command through the motion control algorithm (the regular control and the adaptive weighted control). And the robot can navigate by itself under an unknown and dynamic environment.

Video: nav-video

Prerequisites

1. Window 10

2. Python 3.X

3. TensorFlow-GPU 1.10

4. OpenCv 3.4.X

5. DataBase: MySQL

To store the encoder data (pose, velocity)

Contact us

For any issues, please feel free to contact Teng Ran: rantengsky@163.com

Citation

BibTex

@article{RAN2021389,
title = {Scene perception based visual navigation of mobile robot in indoor environment},
journal = {ISA Transactions},
volume = {109},
pages = {389-400},
year = {2021},
issn = {0019-0578},
doi = {https://doi.org/10.1016/j.isatra.2020.10.023},
url = {https://www.sciencedirect.com/science/article/pii/S0019057820304183},
author = {T. Ran and L. Yuan and J.b. Zhang}}

About

Indoor Navigation for a Differential-Drive Robot in an Unknown and Dynamic Environments

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages