catree / dre_slam

RGB-D Encoder SLAM for a Differential-Drive Robot in Dynamic Environments

DRE-SLAM

Dynamic RGB-D Encoder SLAM for a Differential-Drive Robot

Authors: Dongsheng Yang, Shusheng Bi, Wei Wang, Chang Yuan, Wei Wang, Xianyu Qi, and Yueri Cai

DRE-SLAM is developed for a differential-drive robot that runs in dynamic indoor scenarios. It takes the information of an RGB-D camera and two wheel-encoders as inputs. The outputs are the 2D pose of the robot and a static background OctoMap.

Video: Youtube or Dropbox or Pan.Baidu

Paper: DRE-SLAM: Dynamic RGB-D Encoder SLAM for a Differential-Drive Robot, Dongsheng Yang, Shusheng Bi, Wei Wang, Chang Yuan, Wei Wang, Xianyu Qi, and Yueri Cai. (Remote Sensing, 2019) PDF, WEB

Prerequisites

1. Ubuntu 16.04

2. ROS Kinetic

Follow the instructions in: http://wiki.ros.org/kinetic/Installation/Ubuntu

3. ROS pacakges

sudo apt-get install ros-kinetic-cv-bridge ros-kinetic-tf ros-kinetic-message-filters ros-kinetic-image-transport ros-kinetic-octomap ros-kinetic-octomap-msgs ros-kinetic-octomap-ros ros-kinetic-octomap-rviz-plugins ros-kinetic-octomap-server ros-kinetic-pcl-ros ros-kinetic-pcl-msgs ros-kinetic-pcl-conversions ros-kinetic-geometry-msgs

4. OpenCV 4.0

We use the YOLOv3 implemented in OpenCV 4.0.

Follow the instructions in: https://opencv.org/opencv-4-0-0.html

5. Ceres

Follow the instructions in: http://www.ceres-solver.org/installation.html

Build DRE-SLAM

1. Clone the repository

cd ~/catkin_ws/src
git clone https://github.com/ydsf16/dre_slam.git

2. Build DBow2

cd dre_slam/third_party/DBoW2
mkdir build
cd build
cmake ..
make -j4

3. Build Sophus

cd ../../Sophus
mkdir build
cd build
cmake ..
make -j4

4. Build object detector

cd ../../../object_detector
mkdir build
cd build
cmake ..
make -j4

5. Download the YOLOv3 model

cd ../../config
mkdir yolov3
cd yolov3
wget https://pjreddie.com/media/files/yolov3.weights
wget https://github.com/pjreddie/darknet/blob/master/cfg/yolov3.cfg?raw=true -O ./yolov3.cfg
wget https://github.com/pjreddie/darknet/blob/master/data/coco.names?raw=true -O ./coco.names

6. Catkin_make

cd ~/catkin_ws
catkin_make
source ~/catkin_ws/devel/setup.bash

Example

Dataset

We collected several data sequences in our lab using our Redbot robot. The dataset is available at Pan.Baidu or Dropbox.

Run

1. Open a terminal and launch dre_slam

roslaunch dre_slam comparative_test.launch

2. Open a terminal and play one rosbag

rosbag play <bag_name>.bag

Run on your own robot

You need to do three things:

1. Calibrate the intrinsic parameter of the camera, the robot odometry parameter, and the rigid transformation from the camera to the robot.

2. Prepare a parameter configuration file, refer to the config folder.

3. Prepare a launch file, refer to the launch folder.

Contact us

For any issues, please feel free to contact Dongsheng Yang: [email protected]