Dataset Aracati 2017

This project provides ROS nodes to support the play dataset ARACATI 2017.

The data was collected on the marina of Yacht Club of Rio Grande - Brazil by a Remote Operated Vehicle LBV 300-5 from Seabotix using a forward looking sonar Blue View P900-130 from Teledyne BlueView. A floating board holds the vehicle underwater and a DGPS and a GoPro on the surface. The vehicle drives the floating board and the GPS records the ground truth localization of the vehicle. The data also has a compass, USBL localization, and surface images from GoPro.

The sonar maximum range was set to 50 meters it covers 130 degrees forward the vehicle.

Preview of the dataset available at:

Instructions

This project must be cloned on the src folder of your ROS workspace.

1. Access src folder of your workspace:

$ cd {path to your workspace}/src/

2. Clone this repository on the src folder of your ROS workspace:

$ git clone https://github.com/matheusbg8/aracati2017.git

3. Download the bag file of dataset ARACATI 2017 on this address and save on folder bags of this repository.

4. Build your workspace:

$ catkin build aracati2017

5. Source your workspace again to update the package:

$ source ../devel/setup.bash

If you set your ros workspace on bashrc, you can just close and re-open your terminal.

6. Run the dataset:

$ roslaunch aracati2017 run.launch

Topics

• /aerial_img - A compressed aerial image with the sonar field of view and vehicle path.
• /cmd_vel - Command of velocity of the vehicle, the angular velocity in Z (heading) is estimated from the vehicle compass.
• /dgps - DGPS measurements
• /initialpose - Initial vehicle position. Published once on start of node /odom.
• /odom_pose - Vehicle pose estimated from topic /cmd_vel by node /odom.
• /pose_gt - DGPS position in meters regarding the first vehicle position.
• /son/compressed - Acoustic images from P900-130.
• /son_aerial - Sonar Field of View on aerial image publish by node /aerial_image.
• /surface/compressed - GoPro images on surface.
• /usbl - USBL measurements. When the ping fails the transponder position is published.
• /usbl_point - Vehicle position from USBL in meters regarding vehicle first position.

ROS Nodes

The run.launch file runs the following two nodes:

• odom: Provides dead reckoning localization from the /cmd_vel topic .

  • Subscribes on topic "/cmd_vel"
  • Advirtise topic "/odom_pose"

• aerial_image: Publish aerial images and plot vehicle path

  • Subscribes on topoics: /odom_pose and /pose_gt
  • Advertise topics: /aerial_image and /son_aerial

Thanks

Special thanks to Joel Gaya, Hendry Chame, Eglen Protas, Lucas Longaray, Hallan Neves, José Bratti for helping during the experiment, and professors Dr. Silvia Botelho, Dr. Paulo Drews-Jr, and Dr. Glauber Gonçalves for supporting and supervising.

This study was done by researchers at Federal University of Rio Grande and it was partly supported by CNPq and the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001.