Agenda

Presentation Details

Open Source Robotics at JSK/Tokyo

Kei Okada

Robotics, specially Humanoids research requires long-term vision and efforts due to it’s complexity and synthesis. This talk will show research history on humanoid robots last 20 years at JSK/U-Tokyo and discuss how we have been integrating and maintaining integrated humanoid software over different generation of both students and robots. Rapid-prototyping and contentious-development is the key of these environment and how these features are utilized in national/international projects and challenges. From this perspective, emergence of opensource robotics had great impact since we already had great deal of in-house software resources and had to cope with how to fuse and connect existing internal software and external open software. I’ll not only explain strategy to this integration but also show our automatic translation system of different software component. Lastly, I’ll explain our effort on commercializing our software in Tokyo to expand open robotics software not only academia but also robotic industries.

Robot Middleware Standard: RT-Middleware: Its architecture, implementation and standardization

Noriaki Ando

RT-Middleware (RTM) is an OMG standard specification of a component model and its implementations for robotic systems. We already have a dozen of implementations, and interoperability among implementations are guaranteed by the standard specification. Moreover its ports mechanism enables interoperability to other middlewares such as ROS DDS and so on, and the execution context feature of RTM realizes real-time capabilities, which is utilized in HRP, Hiro/Nextage and SHAFT’s humanoid and so on. In this talk, the concept of RT-Middleware, architecture and implementations will be introduced. New standardization activity in OMG to define unified software component model will also be presented. As other topic, our collaboration with US DoD in DARPA DRC and new projects will be mentioned.

ROS at CSIRO

Paulo Borges

In this talk we will discuss how our team at CSIRO is using ROS in our robotic platforms, including internal projects and development processes.

CSIRO is Australia’s national research agency and our lab, the Autonomous Systems Laboratory, has a very strong focus on field robots, carrying out research on ground, air, and underwater vehicles.

The ROS-based platforms we will present include autonomous industrial ground vehicles, air vehicles, thermal-camera based 3D mapping and the Zebedee 3D laser-scanner.

I will show how the platforms combine out-of-the-box ROS packages with in-house developed packages, indicating where ROS works perfectly as well as some of the issues we face.

Finally, we will also contrast ROS with DDX (Dynamic Data Exchange), which was our successful alternative before ROS was made available.

ROS, Atlas and the DARPA Robotics Challenge

Wyatt Newman and Chris Swetenham

This presentation will review use of ROS in controlling a Boston Dynamics “Atlas” humanoid robot competing in the DARPA Robotics Challenge as “Team HKU” (Hong Kong University). Atlas robots were delivered in August, and the DRC “trials” were held in December, allowing very little time for interfacing to and programming of this new robot. Additionally, a challenging constraint of the DRC is limited communications between the robot in the field and a remote operator. A specific ROS challenge was the need to run local and remote ROS cores to limit communications via a “net tunnel.” Use of ROS will also be described in the contexts of locomotion planning, manipulation planning, vision-guided manipulation, balancing and force control. Live Atlas demonstrations will be performed.

Getting Into ROS: What’s available and how to use it easily

Daniel Stonier

In this presentation I will discuss various approaches and technologies that are at your disposal when wishing to connect to a ROS robot. ROS has a large and sometimes bewildering variety of solutions in its arsenal – ssh tunnels, remotely networked ROS pc’s, rqt plugins, robot web tools, java-android apps, pairing with application managers and also some more esoteric multimastering techniques. However, what is right for your team and your user’s needs? This can be a complex question and a lack of documentation or knowledge about the technologies and languages they use often makes the answer non-trivial.

Here I will introduce each method with some tips and a list of pros and cons. Finally I will conclude with an overview and a few directional pointers so that you can get your own connectivity needs kickstarted without undue time consuming experimentation.