Gazebo & RViz simulation

This assignment consists in a robot simulation based on ROS environment and in c++ language and Jupyter Notebook. Gazibo and Rviz environment has been downloaded from Prof. Carmine Recchiuto's repository for Robotics Engineering students of UNIGE.

Some parameters has been already set in slam_gmapping package needed for this simulation.

Installing and running

At first, install the above package in your ROS workspace by executing in terminal:

git clone https://github.com/CarmineD8/slam_gmapping

Since ROS noetic is used, make sure to also have the 'ros navigation stack' installed by executing:

apt-get install ros-noetic-navigation

Switch on the noetic branch of the packeges by executing:

git checkout noetic

At last, download this package and build the workspace by executing catkin_make command.

Now, to run this version of the simulation digit respectively in three different terminals:

roslaunch final_assignment simulation_gmapping.launch
roslaunch final_assignment move_base.launch
roslaunch final_assignment interface.launch

Gazebo and RViz environment is now available. Otherwise, just run the first two lines and the Notebook to start the interface. By using it, every node run with interface.launch is not needed anymore and, thus, no need to execute it.

Assignment

For this assignment a software architecture for the control of the robot is developed. This architecture is defined as an user interface on Jupyter Notebook that takes user requests and then execute the corresponding controlling mode. There are three different robot control modalities:

• autonomously reach a position in the environment passed as input by the user
• the user takes full control of the robot and drives it with the keyboard
• the user takes control of the robot, but it is assisted to avoid collisions

This interface is developed as three nodes that subscribes and publish to different topics and it implements different functions for every modality. The 'interface_node' manages directly the user interface, 'callbacks_node' calls every time the callback functions of the subscribers and 'driver_node' control the robot both with respect to the minimum distances from the obstacles and to user inputs in user-controlling modalitites. This nodes communicate with custom service.

Here you can find the corresponding documentation for the controlling nodes and the custom service: https://aurorad-hub.github.io/RT2---assignment/

Notebook errors

The Notebook isn't able to drive the robot by only using move_base even if it publishes to move_base/goal topic every time the user decide to send a new target by click on the corresponding button widget. This implies that also position of new targets isn't plotted in the matplotlib widget. Moreover, also plots of robot position and laser scanner data aren't constantly updated.

Simulation topics used

Reach new goal

By publishing in move_base/goal topic, the user can give (x,y) position of the target he wants the robot to reach. For this purpouse the user need to access to the following fields of the topic:

• goal
• target_pose
• pose
• position

And then the following fields can be modify:

• x,y: target new coordinates

This topic type is move_base_msgs/MoveBaseActionGoal.

Goal status

By subscribing to move_base/status topic, the node has access to many fields among which status_list is used to get information about the state of the goal. In this boolean array two elements are important:

• SUCCEDED: is 1 whenever the goal is satisfied
• REJECTED: is 1 whenever the goal is unreachable

This topic type is actionlib_msgs/GoalStatusArray.

Cancel goal

By publishing in move_base/cancel topic, the current goal can be cancelled both by the user and if the goal is unreachable. It only needs to publish an empty request.

This topic type is actionlib_msgs/GoalID.

Scan obstacles

By subscribing to the scan topic, the node has access to many fields among which rages is used to get the information needed: it is an array of 721 elements which contains the distances from the nearest obstacles in a [0 180]° vision range.

This topic type is sensor_msgs/LaserScan.

Velocity

By publishing into the /cmd_vel topic, the node can modify its fields:

• linear: linear velocity array
  • x,y: direction of the linear velocity
• angular: angular velocity array
  • z: direction of the angular velocity

This topic type is geometry_msgs/Twist.

Simulation custom service

A custum service /service handles communication between all nodes by passing information about:

• state of the goal: SUCCESSED or REJECTED
• minimum distances from obstacles: min_r, min_c and min_l
• keyboard commands: dir
• control modality: user driving or user assisted driving