Periodic Report Summary 2 - LOCOBOT (Toolkit for building low cost robot co-workers in assembly lines)
Project Context and Objectives:
LOCOBOT has developed a toolkit for low-cost robots built from a set of plug-and-produce kine-matic modules with compliant yet precise actuators and intelligent sensors for man-machine coop-eration. The toolkit will provide higher flexibility, adaptivity and scalability, all of which are required to meet the challenges faced by manufacturing in the 21st century. Key players in the automotive industry, along with the manufacturers of automation modules and components, as well as high-tech SMEs, are supported by a group of high-quality researchers in solving the project’s technical challenges. Production in the automotive industry is faced with the challenge of high numbers of variants. A robot co-worker will substantially enhance efficiency by cooperating with people and greatly reduce the need for heavy lifting. The need to enter the market early, with reasonably priced vehicles, while increasing production, has to be met by new production technologies such as those developed at LOCOBOT.
LOCOBOT goes far beyond most available systems as it is safe, low-cost and tailor-made, comply-ing with the end-user’s need to produce greener, more customised and higher-quality products for their industry. Stemming from its increased flexibility and efficiency, the immediate impact (2-5 years) of LOCOBOT will be about €150 million in savings. This will be 10 times as much in the fol-lowing years, depending on how the production numbers of the e-vehicle evolve. Three major ob-jectives will be addressed:
• Development of a modular plug and produce robotic assistant platform in which the robot will consist of a set of lightweight, compliant kinematic modules built on a mobile platform.
• Reconfiguration of adaptive control for plug-and-produce components to avoid costly repro-gramming and setup procedures for control algorithms and software. Control algorithms need to be adaptive and self-optimising to account for the different kinematic structures, deal with oscillations induced by the mobile platform and achieve precise positioning.
• Intelligent sensing and actuating structures, for which the robot will be equipped with a stereo camera system and audio components to obtain and process audio visual information, so that it can learn to cooperate with human workers.
At a Glance
Call FoF.NMP.2010-1 Plug and Produce components for adaptive control
Start 2010-08-01
Duration 36 month
Costs 5 320 000 €
Funding 3 740 000 € funded by FP7
FP7 Ref. 260101
Type Collaborative project (generic)
•PROFACTOR GMBH COORDINATOR
•AUDI AKTIENGESELLSCHAFT
•HERIOT-WATT UNIVERSITY
•TECHNISCHE HOCHSCHULE INGOLSTADT
•POLITECNICO DI MILANO
•THE UNIVERSITY OF EDINBURGH
•FERROBOTICS COMPLIANT ROBOT TECHNOLOGY GMBH
•FESTO AG & CO KG
•RIDGEBACK S.A.S. DI PAOLO BARATTINI & C.
•VISUAL COMPONENTS OY
Website:
www.locobot.eu
Project Results:
The ultimate goal of LOCOBOT is to provide safe, tailor made low cost systems which are socially accepted by the worker and enable greener, more customized and high quality products, capable of relieving the human worker from part of the tasks till now not automatised, as well easy adaptability. A a framework which allows designing complex systems based on ergonomics and human robot interaction requirements out of a modular robot tool kit was developed in the project. A Specifically focus was given on how individual patterns of interaction affect and evolve the design of a robot co-worker in a manufacturing environment.
The LOCOBOT consists of an autonomous platform with a resilient arm, a self-adapting FINRAY soft gripper (and other plug&produce grippers), and multiple sensors for safety and HRI and a set of software modules.
The project developed HRI and Technical requirements on the basis of three scenarios in a car assembly line. They include physical cooperation and task sharing with the Human worker that has supervisory role. In these demonstrators the LOCOBOT relieves the human from most of the physical work. An intensive evaluation of on human machine interaction was done in several validation tests.
Main results achieved were the definition and also the demonstration of 3 “near to production” scenarios in an automotive assembly line. The LOCOBOT system includes the following modules: Navigation System, Manipulation System (LocoArm), Object-Recognition, Interaction-Interpretation System (vocal and gesture) and the Safety System. This required a model based engineering process and relevant Integration effort.
A. Demonstrator1 - starter presorting: In this scenario the LOCOBOT has to locate a kart containing the starter of a certain model, scan the content, identify the starter, pick it up and bring it to the orange kart. It needs HRI in case of malposition of the part, presence of paper board sheets between different layers of starters, absence of part/kart/rack, failure in identification of the part, unidentifiable obstacle occupying the work space. Currently this work is done completely manually by the Human worker.
B. Demonstrator 2 - Battery pre-picking: In this case car batteries are delivered to the work area on pallets. The LOCOBOT has to identify the pile of the requested battery model, pick up a battery and bring it to a wheeled kart. The HRI is needed for the lifting of the battery handle (so that can be gripped), removing cardboard layers, malposition of the battery etc. Currently this work is done completely manually by the Human worker
C. Demonstrator 3 - Battery mounting: The car batteries are picked in the right sequence from a three lane slide-rack. Then the LOCOBOT moves to the car that is in the assembly, moved by the lifting crane at slow speed. The LOCOBOT has to track the car speed and present the battery at the correct position inside the hood. The battery will be positioned in collaboration with the human worker.
The platform (System) is a highly flexible solution to many problems that now affect the industrial production. For its practical application the industrial requirement of being 100% failure safe with regards to any task performance requires further research and improvement of the different controls and of the HRI modules, on the other side safety controls appears to be satisfactorily implemented
The project has done an intensive dissemination and exploitation task with a lot of PR activities to show the advantage of such a plug&produce robot as Co-worker in the industry. This includes as well 19x Paper in Proceedings of a Conference/Workshop, 1 Peer reviewed publication and 1x Article/Section in an edited book or book series, various Videos and Showcases (Open days in Brussels) and also an Exploitation plan and ides for further cooperation.
Potential Impact:
A final evaluation of the system under industrial production conditions was done at the end of the project. Goal Objective of the Task was the Evaluation of Human machine interaction and communication during the demonstration tests in the three foreseen validation scenarios using direct observation and an Heuristic of 25 items for each of the 3 areas (ergonomy/usability, health and safety, emotional and psychological impacts). Evaluation after the demonstration through Semi structutured interviews (related to er-gonomy/usability, health and safety, emotional and psychological impacts) to the workers and users involved (10 interviews) including the Heuristic items previously defined. Main objective was the demonstration and evaluation of the system in field tests under near industrial production conditions. Therefore, 3 robot co-worker demonstrators will be set up and optimized for complementary scenarios. In an assembly line the tasks can be summarized in different types. Three typical test scenarios including “delivering of parts”, “helping-hand” and “pre-sorting”: The demonstrators are built upon the prototypical system developed in T5.2. The setting up of the systems in a consecutive way should also demonstrate the ability for rapid reconfiguration, flexibility and adaptations of the developed plug-and-produce components. Accompanied evaluation during demonstration will optimize the use of the LOCOBOT components for specific scenarios.
This Evaluation puts the performance metrics of the prototypical Locobot in relation to the bench-marking metrics defined in WP1. The Locobot was evaluated under near industrial conditions dur-ing the testing of the demonstration scenarios as per Task 6.1-6.3 and the social benchmarking metrics where specified using the questionnaire on human machine interaction and communication from Task 6.4. The result show that hardware requirements are mostly fulfilled or can be easily adapted during further development (e.g. area loading, change of arm modules...). Also for recon-figuration and software aspects, performance metrics are very high. Availability of the system in the demonstration scenarios is not yet on the required level, also hardware layout did not completely fulfil the constraints posed by the original industrial scenarios (D1.1). Also, social factors seem to require some improvement using the feedback from the evaluation with the prototype. However, due to the limited integration time and the high complexity of the project, very good and promising values could be reached. Given the research aspect of the project, not all of the benchmarking metrics strongly influenced by industrial requirements could be reached during the project phase. But most importantly, valuable insights into modular low-cost robotics for industrial applications could be gained and the LOCOBOT prototype constitutes an important milestone for the further development of LOCOBOT systems ready for mass production.
List of Websites:
www.locobot.eu
