Skip to content

Latest commit

 

History

History
133 lines (84 loc) · 4.7 KB

README.md

File metadata and controls

133 lines (84 loc) · 4.7 KB

ROS_Wheelchair

This repository is for the same goal of the prototype edited to be suitable for our Smart Wheelchair final product; Indoor Navigation using ROS. It also contains the description, mapping, localization, and navigation code for the wheelchair which takes the movement order from the Gaze Controller.

The final application assumptions of The Indoor Navigation:

  1. The patient's house is scanned, a complete grid-based map is created and saved, then all final destinations are defined on the map through coordinates.
  2. The desired rooms/final goals location are mapped to i.e. letters A, B, C, and D which is sent to ROS system to move the wheelchair to the target place through a path free of obstacles to ensure safety.
  3. While ROS System is run in this mode, the user is asked to the next destination.

Package Descriptions

  • slam: wheelchair description URDF, simulation Gazebo, visualization RViz, and mapping gmapping launch file.
  • move_robot: sending goals to the ROS navigation Stack wheelchair navigation AMCL and Navigation parameters.
  • odom_pub: taking odometry information from encoder ticks on Arduino.
  • teleop_twist_keyboard: keyboard control of the wheelchair

Build

Steps to build "each in a new terminal tab"

mkdir -p ~/catkin_ws/src
cd ~/catkin_ws/
catkin_make
source devel/setup.bash
cd src/
git clone https://github.com/RoboBrain23/ROS_Wheelchair.git
cd ~/catkin_ws/
catkin_make

Run

Mapping

To launch the urdf Wheelchair description model in Gazebo , start to view the simulated model and take information from the robot's camera in RViz , and moving the wheelchair in gazebo by keyboard 

roslaunch slam gazebo.launch
cd catkin_ws/src/slam/rviz 
rviz -d map.rviz
rosrun teleop_twist_keyboard teleop_twist_keyboard.py
  • After setup, run SLAM: 
roslaunch slam gmapping.launch
cd catkin_ws/src/slam/rviz
rviz -d map.rviz

To create a map of your environment, drive the wheelchair around the environment you want to map.

wheelchair_map.mp4

Once you completed mapping the whole environment, save the map:

rosrun map_server map_saver ~/catkin_ws/src/slam/maps/name_of_map

The final scanned in previous demo be like:

final_map

Autonomous Navigation

Autonomous Navigation can be done using the following -- giving the final destination:

1- Graphical User Interface (GUI).

launch the urdf model in Gazebo, and RViz visualization using the saved map 

roslaunch slam gazebo.launch
cd catkin_ws/src/slam/rviz 
rviz -d navigate.rviz

Using amcl package for localization, Run amcl.launch. The amcl algorithm implements Monte Carlo localization for state estimation. We will use amcl with a map built in the previous step. In the second part, we will create our map using the gmapping package and then use the resulting map for localization.

The move_base package implements an action that will attempt to reach it with the wheelchair. This node links a global and local planner to accomplish its global navigation task. The move_base node also maintains two costmaps, one for the global planner and another for a local planner that are used to accomplish navigation tasks. Run move_base.launch.

roslaunch slam amcl.launch map:='name_of_map'
roslaunch slam move_base.launch
wheelchair_nav_gui.mp4
  • Red arrows, in the beginning, indicate the wheelchair probable location in RViz.
  • Set a goal for the wheelchair in RVIZ (Click "2D Nav goal" and pinpoint the desired location and direction on the map).
  • Green line indicates the path planned.
  • The wheelchair avoids obstacles "according to the pre-saved map" and also Dynamic obstacles that are added when it navigates.

2- Sending the goal location coordinates through the move_robot node.

launch the wheelchair model in Gazebo and simulate in RViz 

roslaunch slam gazebo.launch
cd catkin_ws/src/slam/rviz 
rviz -d navigate.rviz

Then run amcl.launch, giving the name of the map you have created and saved. Run move_base.launch to move the wheelchair though the path planned by the global planner to accomplish the global navigation task. Finally send the location coordinates of the gaol to navigation-stack using move_robot node: 

roslaunch slam amcl.launch map:='name_of_map'
roslaunch slam move_base.launch 
rosrun move_robot move_robot
wheelchair_nav_node.mp4