INS/UWB Integrated Positioning

This ROS package is the implementation of the navigation algorithm illustrated in the article An Approach to Robust INS/UWB IntegratedPositioning for Autonomous Indoor Mobile Robots, Liu J, Pu J, Sun L, He Z., Sensors (Basel). 2019;19(4):950. Published 2019 Feb 23. doi:10.3390/s19040950 using Husky platform as benchmark

Dependencies

In order to use this package you should have ROS installed in your system. Additionally this package depends on some standard python libraries and two other packages used for the UWB Gazebo plugin which are gazebosensorplugins and gtech_msgs both by valentinbarral. Script dependencies.sh should set everything you need

chmod +x ./dependencies.sh && ./dependencies.sh

Run the simulation

You should first set all the environmental variables for Gazebo to work correctly by running setup.bash in your workspace root folder

 chmod +x ./setup.bash && ./setup.bash

And then launch the script start_all.sh which will start:

• Gazebo with Husky model in an empty world with UWB antennas
• The navigation node node_sage-husa.py
• rqt_multiplot with rqt_multiplot.xml config in order to monitor the filter behaviour
• Publisher for Husky velocity command in order to keep the robot in a circular uniform motion

chmod +x ./start_all.sh && ./start_all.sh

In order to start visualizing data in rqt plot you should hit the play button on the top right of each graph

Folders

src

Contains Python scripts which performs navigation. In particular node_sage-husa.py sets the node up and shfaf.py contains the filter class.

data

Contains data exported as csv of simulation run with different filter configurations

launch

Contains roslaunch script to spawn husky with IMU and UWB tag (spawn_husky_uwb.launch, description.launch), Husky controls (control.launch), and to launch the empty world with UWB antennas in place (uwb_empty.launch)

models

Contains Husky urdf xacro model

worlds

Contains different world configuration