ROSCon 2016会议PPT和视频链接

   以下内容来自ROSCON 2016链接：http://roscon.ros.org/2016/#program。

Video recording and archiving is provided by the support of

Locations

• Registration: Grand Ballroom 3, Level 3
• General Sessions: Grand Ballrooms 1-2, Level 3
• Coffee Breaks: Foyer and Grand Ballroom 3, Level 3
• Reception: Foyer and Grand Ballroom 3, Level 3
• Lunch: Foyer and Park Ballrooms 1-3, Level 5Note: the exact scheduling of presentations is subject to change.

Day 1, October 8th

(jump to Day 2)

Time	Presenter(s)	Title	Abstract	Resources
8:00	Everyone	Registration open	Please arrive early to allow time to collect your badge and conference bag before the presentations start. We expect there to be a queue for registration on the first day.
9:00	Brian Gerkey	Opening Remarks		pdf video
9:10	Steffi Paepcke, Louise Poubel (OSRF)	What’s new in Gazebo? Upgrading your simulation user experience!	Gazebo is one of the most used simulators in the ROS community. It has been under heavy development for the past few years and its most recent version, Gazebo 7, comes with myriad new tools and features for new and experienced users alike. Recently, Gazebo development has emphasized user-centered design and improved usability. Updates include not only improved GUI tools and documentation for new folks, but also tools that streamline the workflow for experienced users. We explore new features including: Model Editor, Building Editor, apply force tool, logging and playback, model alignment and snap tools, camera angle controls, plotting, introspection and debugging aids, and more.	pdf video
9:50	Deanna Hood, William Woodall (OSRF)	ROS 2 Update	This talk will summarize the progress made since the last ROSCon update in 2015. Summary will include the alphas released during that time, changes to supported implementations, and the roadmap. The talk will also include demonstrations of new features and highlights of our experiences while using ROS 2 in demos and benchmarking.	pdf video
10:30	Everyone	Coffee Break
11:00	Mukunda Bharatheesha (Delft University of Technology), Ruben Burger (Delft University of Technology), Maarten De Vries (Developer, Delft Robotics BV), Gijs van der Hoorn (Delft University of Technology), Martijn Wisse (Delft University of Technology)	Plan to Win with MoveIt! - Lessons learnt from the Amazon Picking Challenge 2016	This presentation will focus on some of the key MoveIt! practices that we (motion planning team of Team Delft) followed for the Amazon Picking Challenge 2016. Particularly, the following points will be highlighted: 1) making appropriate MoveIt! API choices from a large set of options; 2) difficulties faced such as I/O synchronization with trajectories and collision checking with Octomaps and the corresponding solutions; 3) unsolved problems (mostly with robot driver) while planning around the joint limits of the robot; and 4) general recommendations for OMPL planner configurations with MoveIt!.	pdf video
11:20	Matthieu Amy (Renault SAS), Michaël Lauer (LAAS-CNRS), Jean-Charles Fabre (LAAS-CNRS)	Adaptive Fault Tolerance on ROS: A Component-Based Approach	Every system evolves during their operational lifetime. A system that remains dependable when facing changes (new threats, failures, updates) is called resilient. We propose an approach to safety and adaptive fault tolerance taking advantage of Component-Based Software Engineering technologies for tackling a crucial aspect of resilient computing, namely the on-line adaptation of fault tolerance mechanisms. We will show how this approach can be implemented on ROS and explain some implementation details and the result of different experiments to validate the solution. We will also discuss the how we can use checkpointing technologies to make the ROS master crash-tolerant	pdf video
11:40	Niharika Arora (Fetch Robotics)	Robot calibration	Calibration is an essential prerequisite for nearly any robot. We have created a fast, accurate and robot­agnostic calibration system, which calibrates robot geometry in addition to the typical camera intrinsics and/or extrinsics. The system can be used with a variety of feature detectors to update the cost function and uses the CERES optimizer for the convex optimization. The system then creates an updated URDF containing the calibrated parameters. This talk will cover the details of the robot­agnostic robot_calibration package and describe its use in the fetch_calibration package which can calibrate dozens of parameters on a Fetch robot in as little as 3 minutes.	pdf video
12:00	Iñigo Muguruza Goenaga, Alejandro Hernández Cordero, Víctor Mayoral Vilches (Erle Robotics)	The robot_blockly package: programming ROS with blocks	robot_blockly is a ROS package that allows users to create ROS-based algorithms and behaviors, abstracting its complexity using blocks. The aim of the package is to hide the complexity of programming robots via functional blocks. As a rule of thumb, an average PhD student takes 3 weeks to learn ROS. This makes ROS programming not accessible for the great majority. The robot_blockly package aims to simplify the process of using ROS to the point of putting conceptual blocks together.	pdf video
12:20	Everyone	Lunch
13:50	Yoonseok Pyo, Yoshihiro Shibata, Ryuwoon Jung, Taehoon Lim (ROBOTIS)	Introducing the Turtlebot3	The Turtlebot3 has three concepts: a small, extensible, open robot platform. The goal of new Turtlebot is to reduce the size and cost of the platform without reducing capability, functionality, or quality. Firstly, we will provide small size Turtlebot3 that is enough to be transported without a carrying case in a school bag. Secondly, we will offer extensibility for customizing on the mechanical structure with some alternative options: open­source embedded board, the computer, and the sensors. We will present 12 examples of Turtlebot3. Thirdly, we will provide open­source H/W, F/W, and S/W. You can change your Turtlebot3 design as you want.	pdf video
14:10	Amit Moran, Matthew Curfman (Intel - Perceptual Computing)	Introducing Intel RealSense Robotics All-in-one Perception Device	What if creating perceptive robots would have been super easy? Just plug and play. What would you do? Intel® RealSense™ Technology is a combination of hardware and software for creating innovation around machine perception. Among other things, the hardware includes depth camera and the software includes a set of CV algorithms such as person tracking, scene capturing, object recognition… In this talk we will introduce an innovative All-in-one Robotics Perception Device developed in our innovation lab which can boost the development of any robotic project and enable robots’ perception, with virtually zero installation.	pdf video
14:30	Víctor Mayoral Vilches, Alejandro Hernández Cordero, Iñigo Muguruza Goenaga, Irati Zamalloa Ugarte, Lander Usategui Ugarte (Erle Robotics)	Introducing H-ROS, the Hardware Robot Operating System	H-ROS stands for Hardware Robot Operating System, a DARPA-funded project with the aim of creating easily reusable and r econfigurable robot hardware components to simplify the creation o f robots and allow parts from different manufacturers to interoperate using ROS 2 as its core basis. H-ROS will help companies create robot components that speak ROS natively. With H-ROS, building a robot is about placing H-ROS-compatible hardware components together. Handling robots won’t be restricted to a few with high technical skills but will instead be extended to a great majority with a general understanding of the different H-ROS part types.	pdf video
14:50	Everyone	Lightning Talks I		Videos 1 2 3 4 5 6 7 8 9 10 11 12 13 14
15:35	Everyone	Birds of a Feather I
16:05	Everyone	Coffee Break (and continuation of Birds of a Feather discussions)
16:35	Mirko Bordignon (Fraunhofer IPA), Shaun Edwards (Southwest Research Institute), Paul Evans (Southwest Research Institute), Clay Flannigan (Southwest Research Institute), Gijs van der Hoorn (Delft University of Technology), Paul Hvass (Southwest Research Institute), Ulrich Reiser (Fraunhofer IPA), Nicolas Yeo (ARTC – A*STAR), Min Ling Chan (ARTC – A*STAR)	ROS-Industrial turns four and expands worldwide	Four years after it was first launched by Shaun Edwards, ROS-Industrial is now a worldwide effort to expand and improve ROS adoption in manufacturing environments and industrial equipment. As an initiative, it is supported financially by OEMs and system integrators, whose interests are represented and requests are collected within three ROS-Industrial Consortia managed by non-profit, applied research institutions. As a software platform, it is developed by a free and open community gathering in monthly meetings and operating through a federated development model, much like the ROS developers community. The talk will showcase real-world examples and current efforts to expand the initiative and to continue addressing technical and non-technical concerns.	pdf video
16:55	Levi Armstrong (Southwest Research Institute)	ROS Qt Creator Project Manager Plug-in	The ROS Qt Creator Plug-in is developed specifically for ROS to increase developers’ efficiency by simplifying tasks and creating a centralized location for ROS tools. Since it is built on top of the Qt Creator platform, users have access to all of its existing features like: syntax highlighting, editors (C++, Python, etc.), code completion, version control (Git, Subversion, etc.), debuggers (GDB, CDB, LLDB, etc.), and much more. The talk will cover: a description and motivation; overview of current and future features; and example of how to use the plug-in to manage a ROS Workspace.	pdf video
17:15	Trygve Thomessen, Laszlo Nagy (PPM AS)	The Intuitive ROS UI: FlexGui 4.0 – introduction and industrial applications	FlexGui 4.0 is based upon popular web technologies: HTML5, CSS and JavaScript. This way it is possible to run FlexGUI 4.0 on PC, Android, iPhone, Windows Phone and generally on every device with a modern browser, you will h ave exactly the same user experience on each of them. FlexGui 4.0 communicates using ROS, our choice of middleware for Industry 4.0, IoT. Join the session and let us introduce ourselves, see some of our industrial applications, ask us how could we help you with yours!	pdf video
17:35	Paul Hvass, Shaun Edwards (Southwest Research Institute)	ROS-‘X’ – Focused Initiatives	Have you heard of the “ROS-I” or “ROS-M” projects? What about others? As ROS gains acceptance in various commercial markets, ROS-‘X’ projects seem to organize around them. Such markets include industrial, military, agriculture, aerial, and automotive. An overview of such efforts, some more mature than others, will be presented. The various reasons these projects form will be discussed as well as the impact on the broader ROS community. There is great potential in specialized ROS-‘X’ projects, but these projects also have great responsibility to ensure that the broader ROS project continues and is well supported.	pdf video
18:00	Everyone	Reception
20:00		Fin

Day 2, October 9th

(jump to Day 1)

Time	Presenter(s)	Title	Abstract	Resources
8:30	Everyone	Registration open
9:00	Péter Fankhauser (ETH Zurich), Christian Gehring (ETH Zurich), Remo Diethelm (ETH Zurich), Samuel Bachmann (ETH Zurich), Dario Bellicoso (ETH Zurich), Martin Wermelinger (ETH Zurich), Vasilios Tsounis (ETH Zurich), Michael Bloesch (ETH Zurich), Philipp Leemann (ETH Zurich), Gabriel Hottiger (ETH Zurich), Andreas Lauber (ETH Zurich), Dominik Jud (ETH Zurich), Ralf Kaestner (IBM Research Zurich), Mark Hoepflinger (armasuisse), Marco Hutter (ETH Zurich)	ANYmal at the ARGOS Challenge: Tools and Experiences from the Autonomous Inspection of Oil & Gas Sites with a Legged Robot	At the ARGOS Challenge, we participate with our legged robot ANYmal in the autonomous inspection of oil & gas sites. In this presentation, we share our experience with using ROS in our work and introduce our tools, many of which are available open-source. ANYmal is an electrically driven quadrupedal robot, capable of dynamic running, climbing over obstacles, and scaling stairs. It uses onboard laser range sensors to precisely localize and map its surrounding, enabling the autonomous navigation in difficult environments. For ARGOS, ANYmal carries several inspection sensors and uses computer vision to monitor the state of the site.	pdf video
9:40	Dirk Thomas (OSRF)	The ROS build farm - what it can do for me	The presentation is focused on the Python package ros_buildfarm. The various features provided are being described in detail which will improve the understand of the development process of ROS packages ranging from development and continuous integration over the release process to documentation. It will help you to leverage the provided infrastructure of build.ros.org for the development of your ROS packages and improve your workflow. Additionally it is being demonstrated how the software can be applied locally or e.g. via Travis to run your own “jobs”.	pdf video
10:20	Mike Purvis (Clearpath Robotics)	Robust Deployment with ROS Bundles	Late in 2015, Clearpath was facing a deployment crisis, with software in various states going to customer sites, off-site demos, in-house demos & testing, and developer workstations. There was a build-up of cultural issues and technical limitations in the existing tooling. The solution to all of these issues has been to build the entirety of our robot’s software in a single large workspace (using catkin_tools), and then ship the whole thing as one “fat” deb package. Lessons learned will be presented, along with a brief example of building a customized desktop_full bundle for Ubuntu - this should be sufficient to kickstart the efforts of anyone else who’d like to set up a similar build. The demonstration will highlight in particular our contributions to upstream ROS tooling which have been made in the course of this development work, hopefully merged in coming months	pdf video
10:40	Everyone	Coffee Break
11:10	Ian Chen (OSRF)	Cloudy with a Chance of Simulation	The DARPA Virtual Robotics Challenge gave birth to the first version of CloudSim. It provided a centralized platform that facilitated teams of participants from around the world to compete in simulation simultaneously. Since then, various projects have demanded the need for a similar cloud hosted environment for robotics competitions. In this presentation, we reveal a new, redesigned CloudSim that aims to be a generic tool for running robotics software and simulations on the cloud. We present some of its key features including simulation instance sharing, reusable front-end web components, a WebGL visualization client, and security and authentication.	pdf video
11:30	Ricardo Tellez (The Construct)	Robotics Benchmarking with ROS	Benchmarking in robotics is an unsolved issue. Having a benchmarking system for robotics would speed up development of robotics and research since results may be compared against each other in same conditions, and new robotics works may finally build build on previous results. The purpose of this session is to describe a framework for a universal benchmarking system for robotics based on Gazebo simulations, where ROS has the central role of becoming the standard of communication between hardware and software.	video
11:50	Alex Henning (Fetch Robotics)	Physical Continuous Integration — CI on Real Robots!	Fetch Robotics has developed a continuous integration (CI) system for testing new code in a complete system including multiple real robots. Multiple ceiling­mounted cameras and on­robot rosbag recordings allow developers to dig in and diagnose bugs when found. This talk will cover the tools and methods we have used to deploy and monitor tests on our fleet of robots. It will also cover additional considerations necessary for running real world tests and include a case study of using physical continuous integration to verify bug fixes to the ROS navigation stack.	pdf video
12:10	Everyone	Lunch
13:40	Ruffin White (University of California, San Diego), Morgan Quigley (OSRF)	{,S}ROS: Securing ROS over the wire, in the graph, and through the kernel	This talk will demonstrate the uses of SROS, a proposed addition to the ROS API and ecosystem to support modern cryptography and security measures. While detailing current progress and further proposals, a working demonstration of each major advancement will be shown, including: over­the­wire cryptography for all messages, services and parameter transport, namespaced access control enforcing graph policies/restrictions, and finally process profiles using Linux Security Modules to harden a node’s file, device, signal, and networking access on host machines. You’ll gain a greater awareness of the vulnerabilities in ROS and ongoing developments to secure robotics subsystems for the future.	pdf video
14:00	Adam Dąbrowski, Rafał Kozik, Mateusz Macias (PIAP)	Evaluating the resilience of ROS2 communication layer	One of the major reasons for development of ROS2 is the migration of communication layer to DDS. Robot users around the world, hitherto discouraged by the ROS shortcomings, might finally be considering joining the community. For all those who see the promise, there is a question of whether ROS2 will deliver on it. We introduce a project for measurable, repeatable evaluation of ROS2 communication in less than perfect networks. We will present the most recent results from the evaluation tool and draw conclusions: what has been achieved? What are the most important gaps to address?	pdf video
14:20	Jan Carstensen, Axel Rauschenberger (Leibniz Universität Hannover)	RTROS – A real-time extension to the Robot Operating System	RTROS enables ROS users to easily port or develop ROS nodes using many of the benefits from the ROS-ecosystem, while satisfying hard real-time constraints. Tested for Xenomai and preempt-RT the library offers deterministic communication and core functionality to develop a real-time robotic system. RTROS is designed such that standalone RTROS or ROS/RTROS system are possible. Thus, the use of standard ROS nodes for uncritical nodes is still possible.	pdf video
14:40	Everyone	Lightning Talks II		Videos 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
15:25	Everyone	Birds of a Feather II
15:55	Everyone	Coffee Break (and continuation of Birds of a Feather discussions)
16:25	Dejan Pangercic, David Ball, Markus Hoeferlin, Daniel Di Marco, Steffen Fuchs, Tobias Mugele, Tobias Dipper, Gabriel Gaessler, Andreas Michaels, Sandeep Bairampalli, Maurice Gohlke, Roland-Kosta Tschakarow, Amos Albert(Deepfield Robotics)	Agricultural Robotics with ROS at Bosch: From the internet of fields to the internet of plants	Bosch Startup Deepfield Robotics is creating innovative systems for automated plant phenotyping. Our system comprises semi-autonomous robots for data acquisition and a processing pipeline for detection and analysis of plants. Our system is based extensively on ROS. This talk will provide a detailed description of our complete system focusing on our phenotyping plant counting pipeline, the robot’s software and real-time design, our best practices for ROS bagfile management, life cycle management, build farm, provisioning, within the ROS framework. This talk will be of interest to other startups using ROS and other groups wishing to increase their level of quality. So far we have processed 27 terabytes of data and detected over 2 million plants across multiple days.	pdf video
16:45	Dejan Pangercic (Deepfield Robotics), Daniel Di Marco (Deepfield Robotics), Florian Friesdorf (Ternaris UG), Marko Durkovic (Ternaris UG)	Bagbunker - Tool for Large Data Storage, Analysis, Viewing and Testing	Managing large datasets is a challenging, but often under­appreciated task for roboticists, especially in outdoor applications such as agriculture, underwater and space robotics, as well as autonomous driving. We have released bagbunker, a web­based tool for storage, management and retrieval of large datasets such as rosbags. Bagbunker is like a photo app, just with a different data. It is specifically designed to improve the task of managing large datasets, and also for engineers not trained in database systems. In this way, it greatly speeds up development and debugging by keeping the data under central control and providing a web­based interface to filter, inspect, visualize and interact with the data. In this hands­on session participants will learn how to efficiently manage their large datasets, including how to search and retrieve rosbags and how to write jobs to analyse rosbags. Additionally we will also present how to use bagbunker for an image classification regression test.	video
17:05	Norman LI, Paul Yang (DJI)	A robust flying platform for ROS developers	The presentation introduces an accurate, stable and reliable flying platform DJI Matrice 100 together with its corresponding ROS packages to aero robotics developers. Developers can monitor and control aircraft flight behavior with the help of On board SDK ROS packages, while utilizing the built-in Intelligent Navigation Modes to create autonomous flight paths and maneuvers. Its ROS packages are powerful and easy to use, which have helped many Matrice 100 developers earn great achievement in their research and development process.	pdf video
17:25	Ryan Gariepy	Closing Remarks		pdf video
17:35		Fin