In this episode I talk to Florian Lier about what his team ToBI is doing to program Pepper the robot for the Robocup@Home competition.
We will explain the whole architecture they are using to program Pepper with ROS for the competition including:
In which computer to run which code
ROS wrappers for NaoQi
How to use a good Pepper simulation using Morse simulator
Which environments to use to develop robot behaviors, and where to find them
How they are integrating with automatic regression tests
Selected quote from the interview:
We are using a meta-build system that has templates for ROS, combined with some recipes on which you define the type of software you want to deploy. You put the recipes into a system description that is what you deploy into the ICI servers.
In this episode I talk to one of the main developers of the Gazebo simulator: Louise Poubel. She tells us about the ins and outs of Gazebo, how do they organize at OSRF to develop the simulator, why to discard Gazebo 2 and how to install the newest versions. She will also clarify us what are the Ignition libraries and why OSRF is putting so much attention to them for the development of Gazebo. Additionally, she will talk about Gazebo plugins, SDF versus URDF, rendering in Gazebo, skeleton animation and many more things.
Selected quote to from the interview
There is always a compromise in where we use our resources for developing Gazebo. At present we are putting a lot of attention to rendering, and Gazebo 9 comes with a camera lens flare (there’s a nice gif here). We are adding those quality details that make the rendering much better.
In this episode, Jack Pien will talk about how to use accessible robots based on ROS and OpenCV to learn robotics. Those robots are called ROSbots and Jack is the creator of them.
ROSbots come with ROS and OpenCV installed on a Raspberry Pi onboard the robot, ready for the user to create applications for it using the camera and the motor.
The purpose of those robots is to provide hackable robots, on which you can put your hands on, and implement all those robotics algorithms that you study.
Selected quote to from the interview
The vision with the ROSbots is to have a set of software bundles/lessons that you can use to run academic concepts of robotics on the robot
In this session of The ROS Developers Podcast, I chat with Cecilio Angulo, associate professor at Technical University of Catalonia and director of the Intelligent Data Science and Artificial Intelligence Research Center. We will talk about a project he started to bring Aibo robot back to life again by using ROS.
He talks about how the whole thing started and why. Then he describes the different parts implemented in the Aibo ROSification and how the Aibo simulation in Gazebo is important for the development. He also describes the application that makes several Aibo imitate to each other by connecting all of them through a ROS node, as well as the modules for accessing different sensors of Aibo. Finally, we talk about how to make a ROSified Aibo reproduce the famous MTN (motion) files that Sone generated containing predefined movements.
Selected quote to from the interview
By ROSifying the Aibo robot, you can apply reinforcement learning, dynamic motion primitives, deal with all the images using deep learning… all the ROS world is there to be used with the robot.
In this episode of The ROS Developers Podcast, I chat with Carlos Rosales, CTO of Beta Robots a company dedicated to apply ROS into real life industrial projects.
He talks about the simulation model of the softhand, an open source hand with 5 fingers that is flexible and that absorbs collisions. He will also talk about different projects he and his company have been involved including industrial manipulators, perception and grasping. Finally he will provide some information about the Master of Robotics he is doing, including some links to the git of the master and the slides.
Selected quote to from the interview
If you open Kdevelop IDE from a ROS configured terminal, it will provide you with ROS definitions and autocompletion.
In this episode, I interview Enrico Mingo Hoffman, postdoc researcher at the Humanoids & Human Centered Mechatronics Lab of the Italian Institute of Technology. He will talk about their open implementation of the Stack of Tasks for the whole body control of humanoid (and non-humanoid) robots. He will also explain how their development environment at the lab works, and will discuss some differences between ROS and YARP, another framework for robot programming developed at Italy. He will end suggesting about the creation of an OSRF – Europe. Do not miss that!
Selected quote to from the interview
What I expect at the beginning is that the robot stands… without exploding!
