Community Research and Development Information Service - CORDIS

FP7

FACTORY-IN-A-DAY Report Summary

Project ID: 609206
Funded under: FP7-NMP
Country: Netherlands

Periodic Report Summary 2 - FACTORY-IN-A-DAY (Factory-in-a-day)

Project Context and Objectives:
Factory-in-a-Day aims at improving the competitiveness of European manufacturing SMEs by removing the primary obstacle for robot automation; installation time and installation cost. The high costs result in payback periods, making the investment in robotized automation economically unattractive. Factory-in-a-Day will reduce the installation time (and the related cost) from months to one single day. The project follows the steps of such an installation day.
(1) New standardized arms and hands are combined with 3D printed custom parts to be designed in a matter of hours from novel design templates.
(2) Robots are placed in an unaltered production location, where new self-calibration routines and a novel software framework allow effortless interconnection of robot components and existing machinery.
(3) For selected application domains (e.g. mould finishing and assembly) a set of novel learnable skills (cf. "Apps") allow rapid teaching of the production tasks.
(4) The robots collaborate safely with humans in a shared un-fenced workspace due to safe robot arms with dynamic obstacle avoidance, made possible by novel proximity-sensing skin and online path re-planning algorithms. Augmented reality lets the robots project their intended motion plans to inform the workers.
(5) Finally, Factory-in-a-Day also innovates the organizational aspects of robot installation.
The project is driven by Europe’s top robotics researchers (e.g. TUD, KUL, TUM, Fraunhofer) and industry players (Philips, Universal Robotics, Siemens, Materialise). The second-largest temp agency Randstad and its participating client SMEs will test hybrid human-robot teams performing short-batch production work. Together with novel certification procedures proposed in Factory-in-a-Day and strong , Europe-wide dissemination targeted at manufacturing industry, the technological and organizational innovations are poised to radically change the robot automation sector.

Project Results:
Since the beginning of the project up to the second reporting period (Month 36), significant progress has been made in all Work Packages. After setting up a strong management procedure (which has led to on-time delivery of most results) in WP1, we have analysed the new Factory-in-a-Day business models in WP2, worked on ultra-adaptive grippers made with 3D-printing technology in WP3, created skin sensors and sensing algorithms in WP4 and a skill learning language in WP5, which was all brought together with the rapid deployment software of WP6 into first demonstrations in WP7. All of this has been disseminated through the website and a large number of presentations in WP8. The highlights are the following.

In the first reporting period (month 1-18), developments in WP2 were very rapid, resulting in tools such as a QuickScan reported in Deliverable 2.1, and a human factors quick scan in Deliverable 2.2. During the current (second) reporting period, month 19-36, two key results were produced. First, Deliverable 2.3 contains the blueprints for the business concepts and organizational structures needed to successfully implement Factory-in-a-Day results in industry. In essence, the organization of the new type of robot service provider must combine a quick and standardized customer approach with technical flexibility, fast delivery, high level of quality and good personal support. The ideal organization structure is divided in three parts: business, delivery and support. We note that actual implementation of the business concept is now waiting until the Factory-in-a-Day technologies are more mature. Second, Deliverable 2.4 was produced which identifies approaches to speed up the deployment process by certifying parts of the automation system and preparing technical documentation application-independent in advance. The deliverable furthermore identifies the importance for Factory-in-a-Day of the ISO standardisation committee dealing with modularity for service robots.

Workpackage 3 has seen steady progress throughout the project. In the first reporting period, in Deliverable 3.1 the overall Factory-in-a-Day workflow has been presented, and Deliverable 3.3 contained the elaboration of 3D-printing specific part of that same workflow. In the second reporting period, according to schedule three more deliverables have been readied, one with some delay. The first result is Deliverable 3.4, a library of generic 3d printable components, especially aiming at quick development of robot grippers. Second, Deliverable 3.7 describes design templates for mould polishing components. This pertains to one of the demo setups of Factory-in-a-Day aiming at quick installation/setup of an automated polishing robot. Note that we had a project amendment to move this activity away from WP5 into WP3, where (in hindsight) it is more fitting due to the large hardware component of the work. Third, we have almost finished the work on Deliverable 3.2, an instant simulation tool to be used by robot system integrators for very fast (Factory-in-a-Day style) negotiation with customers about the required robot system to be installed.

After the production of a superb skin in the first reporting period within Workpackage 4 (Deliverable 4.1), the work in the second reporting period in T4.1-T4.3, T4.5 and T4.7 has been extensive and strongly integrative between many Factory-in-a-Day partners to use the skin with dynamic motion planning. We are quite proud that our combined efforts led to a double victory in the prestigious Amazon Picking Challenge in July 2016, with underlying technologies developed in Tasks 4.3, T4.4 and T4.6. All these technologies were/are being reported in Deliverables D4.2-D4.4, some with about two months delay but still we are very proud of the technically integrated results obtained.

If WP4 was the “workhorse” on system integration, WP5 clearly was the workhorse on scientific discovery and publication in the second reporting period, boasting eight scientific articles. Building on foundational results from the first reporting period described in Deliverable 5.1 (on novel underlying models and modelling languages for the creation of learnable skills), we have now produced the first fully integrated demo of a “learnable skill”, namely the skill to pick, quality check, sort, and pack oranges. This skill can intuitively be taught by human instructors. In addition, with Deliverable 5.4, partner Universal Robots has delivered URCaps, an open system in which any vendor can add components for easy and quick integration. We are proud to say that the research prototype URCaps has even already resulted in a commercial roll-out as the “Universal Robots+” developer program.

Within Workpackage 6, we have completed the development of an integrated testing/evaluation infrastructure for all types of required testing, from low-level unit tests via simulation and stored-data tests up to full flanged hardware-in-the-loop and application level tests. We also have provided a number of metrics to measure the quality of a new software component. A prototype toolbox for quick robot software integration was presented in Deliverable 6.2 and the description of the final toolbox is given in Deliverable 6.3. The toolbox was used successfully in several demonstration cases for Factory-in-a-Day. Note that the work in WP6 of Factory-in-a-Day has resulted in significant impact through the recent successful H2020 grant application “ROSIN” for further development of the underlying ROS-Industrial software framework and community.

In WP7, during the first reporting period of the project, we defined an extensive list of performance indicators (see Deliverable 7.1) and selected the following five most important ones:
KPI Definition
Cost of Installation This is the overall costs of installation, in EUR, or as a fraction of the total unit value.
Reuse Value Also referred to as ‘leasability’ factor. It captures the level of generic technology.
Installation Time This is the overall time of installation, measured in days.
Reconfiguration Cost of Installation This is the overall costs of installation, in EUR, or as a fraction of the total unit value.
Reconfiguration Installation Time This is the overall time of installation, measured in days.

During the second reporting period, we have executed about a dozen experimental cases and we have measured the score on the above Key Performance Indicators. Throughout the project we have seen a steady increase in the task complexity of these demonstrators, due to which the installation time has not always improved. A core finding is that more focus is needed for the reliability tuning at the end of the installation process, which is crucial for usability but which tends to significantly increase the installation time.

Finally, through a combined effort in WP8 we have presented the project on average 2-3 times per month during the second reporting period, addressing audiences ranging from industry to general public to scientific audiences. The project has a very clear international presence.

Potential Impact:
When Factory-in-a-Day is completed, we expect to have developed and combined all the hardware-, software-, and organizational tools to install factory robots for SME’s within a short time, down to a day for the simpler cases. The collection of tools will have been reported in such a way that the knowledge is available for European-wide roll out. Moreover, various project partners will already have commenced putting the knowledge into practice. The expected impact will be as it was listed in the original proposal:

List of Websites:
http://www.factory-in-a-day.eu/

Contact

Dunja Swierstra, (Project Manager)
Tel.: +31152787880
E-mail
Record Number: 195493 / Last updated on: 2017-03-14
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top