Robocup_at_work

RoboCup@Work is a new competition in RoboCup that targets the use of robots in work-related scenarios. It aims to foster research and development that enables use of innovative mobile robots equipped with advanced manipulators for current and future industrial applications, where robots cooperate with human workers for complex tasks ranging from manufacturing, automation, and parts handling up to general logistics ([], 2017)

media type="youtube" key="38TtlynYd6w" width="560" height="315" align="center"

=Scientific Challenges= RoboCup@Work addresses many of the standard scientific challenges for robotics, including the following non-exhaustive list: In addition, the RoboCup@Work League specifically targets several new challenges, which so far are not pursued by other competitions or RoboCup leagues:
 * Perception in static and dynamic environments under varying environmental conditions
 * Path planning and motion control of mobile bases in dynamic environments
 * Grasp planning, trajectory planning, and motion control of mobile manipulators
 * Planning and decision making
 * Representation of plans, knowledge, strategy and tactics
 * Adaptivity and learning
 * Cooperation in both cooperative and competitive environments
 * Human-robot and robot-robot interaction
 * Design, construction, and operation of robust robots at affordable cost
 * Simulation, evaluation, and benchmarking of advanced robot systems
 * **Mobile Manipulation:** While until the recent past industrial robotics concentrated on highly-precise but non-mobile manipulators, mobile robots either had no manipulators or only low-DoF robot arms. Recently, this situation is changing, and both the research community as well as industry have developed a strong interest in serious and robust mobile manipulators. Introducing a competition that fosters research in that direction comes very timely.
 * **Logistics:** Logistics is an enormously important area in practically every business-related domain. It plays practically no role in any of the established RoboCup leagues so far. Rescue simulation league is a notable exception, but poses quite different constraints on logistics problems than those targeted by RoboCup@Work.
 * **Cooperative Mobile Manipulation:** Once mobile manipulators are not a fantasy any more but reality, the next step would be to have such mobile manipulators cooperate with humans and/or other mobile manipulators. Reaching this objective requires solving numerous additional problems, but also opens many new application scenarios.
 * **Multiagent Planning and Scheduling, and Multi-Criteria Optimization:** The value of classical task planning in soccer and rescue leagues is rather limited; if it used at all, these planners have to deal with constraints that are quite different from those in most industrial domains. Task planning may eventually be of great interest in RoboCup@Home, but so far most teams work with preprogrammed scripts or state machines executing routine activities. In RoboCup@Work, however, multiagent planning and scheduling with multi-criteria optimization is of immediate interest, and has a large potential for innovative applications.

=Industrial Challenges= While the robotics and automation industry is providing excellent technology for automating mass production, increasing competition is forcing also many small and medium-size companies to improve their production processes. These companies often do not have mass production, but have much smaller lot sizes, see frequent changes in the production setup, and generally need much more flexible organization of their production and logistics facilities. Some candidate tasks to be handled by innovative technology like mobile manipulators include: It should be noted that most of these tasks are such that they involve a logistics operations, i.e. moving parts from a usually not precisely specified source location to a not precisely specified target location. Insofar these tasks are quite different from the usual operations of industrial robots in mass production processes. Also, the target object to be manipulated may be underspecified: neither source nor target pose may be known and have to be determined dynamically, only the object class may be known (e.g. a box), but not its dimensions, requiring grasp poses to be determined dynamically. Scenarios for testing the above abilities can be set up to be deterministic initially, but varying degrees of dynamics and stochasticity (e.g. through failures, continuous arrival of new orders, etc.) can be added. Several of the aforementioned tasks bear a large potential for cooperation, being it the cooperation between a parts loading robot and the truck onto which parts are to be loaded, the cooperation of several robots to complete a complex task, or the cooperation between robots and humans, where the robots assist the human e.g. in a complex assembly process.
 * Feeding parts from a palette onto a machine
 * Loading parts from machines onto a palette
 * Feeding a set of parts from containers into a tray for a subsequent assembly process
 * Performing preassembly tasks
 * Packing parts and boxes into containers
 * Unloading boxes from containers
 * Sorting objects from unstructured heaps
 * Wrapping objects (e.g. before packing and shipping them)
 * Assembling complex objects from parts based on an example
 * Painting objects according to an example given as image
 * Moving parts and objects around

more information: [|http://www.robocupatwork.org]