In this ROS LIVE-Class we’re going to learn how to create our own Qlearning training for a cart pole in Gazebo using both OpenAI and ROS.
We will see:
How to create a Python training program that uses OpenAI infrastructure
How to create the environment that allows to get the observations and take the actions on a Gazebo simulation
How to interface everything with ROS
Every Tuesday at 18:00 CET/CEST.
This is a LIVE Class on how to develop with ROS. In Live Classes you practice with me at the same time that I explain, with the provided free ROS material.
IMPORTANT: Remember to be on time for the class because at the beginning of the class we will share the code with the attendants.
IMPORTANT 2: in order to start practicing quickly, we are using the ROS Development Studio for doing the practice. You will need a free account to attend the class. Go to ROS Development Stdio and create an account prior to the class.
[irp posts=”9720″ name=”ROS Developers LIVE-Class #22: How to create an OpenAI environment for your robotic problem”]
In this video we are going to see why a ROS developer can’t launch a PR2 Gazebo simulation, mainly because the ROS controllers are not found.
We will see which are the ROS packages which provide all the required controllers for the PR2. We will also see how you can apply this check to your own robot simulation, in case it complains about the controllers not being loaded.
The above command should output various controllers and thier status.
Now the above case is for the working environment where all the required packages are installed. If some packages are missing then we should systematically look for packages and dependencies. The packages mostly fall into two categories: gazebo-ros controllers and robot specific controllers (pr2-controllers specifically for our case)
ros-control : To list the available ros-control packages use the following command
$ dpkg -l | grep ros-control
There are two important packages in this category ros-kinetic-gazebo-ros-control and ros-kinetic-ros-controllers
gazebo-ros : To list the available gazebo-ros related packages use the following command
$ dpkg -l | grep gazebo-ros
The output of the command should contain the following two names ros-kinetic-gezebo-ros and ros-kinetic-gezebo-ros-control (this one was seen in previous commands output also)
controller interfaces : To list the installed controller interface packages use the command
$ dpkg -l | grep controller-interface
The output of this command should show the following names ros-kinetic-controller-interface, ros-kinetic-controller-manager, ros-kinetic-controller-manager-msgs
controllers : In gazebo we use various controllers to actuate different types of joints. We need to check for the following ones
joint state controllers : The output of the following command should contain ros-kinetic-joint-state-controller
$ dpkg -l | grep joint-state
joint trajectory controllers : The output of the following command should contain ros-kinetic-joint-trajectory-controller
$ dpkg -l | grep joint-trajectory
effort controllers : The output of the following command should contain ros-kinetic-effort-controllers
$ dpkg -l | grep effort-controllers
velocity controllers : The output of the following command should contain ros-kinetic-velocity-controllers
$ dpkg -l | grep velocity-controllers
Finally we need to check for PR2 robot related controllers to be able to launch the simulation. Use the following commands
We expect to see the following names in the output of those commands
ros-kinetic-pr2-controller-manager,
ros-kinetic-pr2-controller-interface,
ros-kinetic-pr2-controller-configuration-gazebo,
ros-kinetic-pr2-controllers-msgs,
ros-kinetic-pr2-mechanism-controllers,
ros-kinetic-pr2-mechanism-diagnostics,
ros-kinetic-pr2-mechanism-models,
ros-kinetic-pr2-mechanism-msgs
If we have all of these packages installed in our computer, we will be able to launch the pr2 robot. That’s all for this post. Thanks for watching the video, please leave your comments and subscribe the channel if you want to see more such videos.
Did you like this video? Do you have questions about what is explained? Whatever the case, please leave a comment on the comments section below, so we can interact and learn from each other.
If you want to learn about other ROS topics, please let us know on the comments area and we will do a video about it.
Basically what we are going to teach in this extra Live-Class is: 1- How to record a trajectory that you would like the robot to do 2- How to make the robot reproduce that trajectory
Everything using the GPS as localization system.
This time we are going to use the ROS Development Studio to do the practice. Please you need to have an account at RDS (free account is ok, and fully working). Please it is very important that attendants have the account created prior to the class so we do not waste time with that process.
All the attendants will receive a full notebook and code that does the two points. The notebook will explain step by step how the code works and where to change for another robot different than the one used in the simulation.
This is an Extra Live Class requested by the ROS Agriculture community (http://rosagriculture.org/), that is why we are doing this Live Class on Saturday (instead of Wednesday). This is an extra class. Wednesday Live Class are still on the schedule (check it out Wednesday schedule here: ROS LIVE-Classes).
In this ROS LIVE-Class we’re going to create a world in the Gazebo simulator for the previous differential drive manipulator we created in the previous class, so the robot can navigate around and interact with the objects.
The model of the robot was created using URDF. However, the model of the environment will be created using SDF.
We will see: ▸ How to create the world for the robot using SDF ▸ How to add models of any object you may think of ▸ How to spawn ROS based robots in the world
Part 1
Part 2
Every Wednesday at 18:00 CET/CEST.
This is a LIVE Class on how to develop with ROS. In Live Classes you practice with me at the same time that I explain, with the provided free ROS material.
IMPORTANT: Remember to be on time for the class because at the beginning of the class we will share the code with the attendants.
IMPORTANT 2: in order to start practicing quickly, we are using the ROS Development Studio for doing the practice. You will need a free account to attend the class. Go to http://rds.theconstructsim.com and create an account prior to the class.
In this ROS LIVE-Class we’re going to show how to create a Gazebo plugin for a robot simulated world. The plugin will allow us to connect/disconnect the light of the world by means of a ROS topic. The plugin will be created using C++.
We will use the simulation we created on the previous Live Class of a wheeled robot on a home room. Do not worry, you will receive the code of the simulation at the beginning of this Live Class (only provided to those who attend live to the class. The code will be provided during the first 5 minutes of the class, so be in time).
We will see: ▸ Which types of Gazebo plugins exist ▸ How to create a world plugin that allows you to control anything in the simulated world (very useful for Reinforcement Learning) ▸ How to add ROS to the plugin so we can use it from our ROS programs.
Every Tuesday at 18:00 CET/CEST. This is a LIVE Class on how to develop with ROS. In Live Classes you practice with the teacher in real time, at the same time that he explains the lesson with the provided free ROS material.
IMPORTANT: Remember to be on time for the class because at the beginning of the class we will share the code with the attendants.
IMPORTANT 2: in order to start practicing quickly, we are using the ROS Development Studio the practice. You will need a free account to attend the class. Go to http://rds.theconstructsim.com and create an account prior to the class. Otherwise you will not be able to practice during the class.
In this Live Class we are going to see how to apply all that we have learnt in the previous classes about Gazebo. We are going to execute ARIAC simulation made by the OSRF. We are going to learn how to manage that simulation in order to make the robots perform useful tasks in the environment. We will indicate where Gazebo plugins were applied, how SDF was used to create the environment, were ROS control was applied, where ROS navigation was applied, etc. We will see how to create a simple Python program that allows us to make the robots do useful things.
A new ROS Live Class every Tuesday at 18:00 CET/CEST. This is a LIVE Class on how to develop with ROS. In Live Classes you practice with me at the same time that I explain, with the provided free ROS material.
IMPORTANT: Remember to be on time for the class because at the beginning of the class we will share the code with the attendants.
IMPORTANT 2: in order to start practicing quickly, we are using the ROS Development Studio for doing the practice. You will need a free account to attend the class. Go to http://rds.theconstructsim.com and create an account prior to the class.
