Automated production lines are central to modern, industrial manufacturing, saving people from the burden of repetitive work. The use of automated production lines is however limited by the cost of creating custom-made machines for difficult tasks that industrial robot arms in turn are too slow to handle efficiently. As an alternative, we are investigating the concept of morphing production lines (MPL) that automatically can change their shape to adapt to the task at hand. This will create a production line that is both flexible and fast.

Contents

edit

News

  • 15/3-2008: First public release of the USSR simulator [1]
  • 21/12: Project staff on well deserved Christmas vacation
  • 10/11: Post doc position filled - all applicants notified
  • 15/9: Meet a large part of the team at IROS 2007 at the Self-Reconfigurable Robots/Systems and Applications workshop [2] or at the Robocomm conference [3]
  • 14/9: Submissions of application for the post doc position closed - applications are currently being evaluated
  • 24/8-2007: We have an open position for a post doc in the project see [4] for further information.
  • 1/8-2007: The Ph.D. position has been filled, a post doc will shortly be announced
edit

Overview

In this project we will develop a prototype of a morphing production line based on the ATRON self-reconfigurable robot. Self-reconfigurable robots are robots built from modules that automatically can change the way they are connected and thereby change the overall shape of the robot. Self-reconfigurable robots have been an active area of research for two decades, although three-dimensional shape change has only recently been demonstrated, for example with the ATRON system. The project is co-financed by the Danish Agency for Science, Technology and Innovation with a budget of 4.5 Million DDK.

As an example of how a morphing production line could be realized using the ATRON self-reconfigurable robot, consider the following mockup:

ATRON production system mockup

The modules form a layer of modules able to transport objects and a robot arm for manipulation. It is important to realize that the morphing production line shown only looks like it is doing the task, but it is in fact not: the modules are in principle physically capable of manipulating the box but lack the necessary sensor information and programming. The goal of this project is to change that by building a new generation of modules and a supporting software platform enabling rapid development of flexible applications.

edit

Current activities

This section is regularly updated to reflect the current activities in the project. The main activites of the project are development of new revision of the ATRON hardware and development of a new software platform supporting self-reconfigurable robots in general and morphing production lines in particular. Nevertheless, we expect that other related activities will also develop as the project progresses, such activities will also be listed here on this page.

Last update: 10/4-2008.

edit

Hardware development

After the initial assessment of the existing ATRON hardware we have decided to pursue two development strands in parallel. We have found that we can integrate SunSpots into ATRON modules and thereby with a relatively little change to the hardware raise the abstraction level of the software API so even computer scientists do not get confused. This solution is only temporary since the module is still handicapped in terms of sensing capabilities and suffer from a number of minor hardware problems, but will allow the software effort to continue at full speed. Details are provided here: Interfacing the Atron modular robot with a SunSPOT In parallel, we will develop a new generation of electronics to fix these hardware issues and according to plan have ATRON II ready towards the end of 2008.

edit

Software assessment

There is no existing repository of software, so as a first step we are looking at existing control approaches as well as software from other domains such as sensor networks. Concretely, we have ported TinyOS to the ATRON robot, which provides vastly improved stability and allows us to make use of numerous features found in middleware for sensor networks.

edit

Development of an improved simulator

The current ATRON-simulator is transition-based meaning that it does not simulate physics but rather simulates a simplified mathematical model of the self-reconfigurable robot. The key idea of morphing production lines is not to shapechange but rather to manipulate other physical objects; to simulate such manipulation friction must also be simulated. Rather than developing yet another ATRON simulator, we have deveoped a unified simulator for self-reconfigurable robots described here: [5].

The simulator now supports the ATRON and Odin modular robots as well as a few others. Although the simulator is written in Java, controllers written in C can be interfaced directly to the simulator, using a generic wrapper written using JNI.

edit

Frameworks

We are currently developing a framework for facilitating programming ATRON modules. The framework focuses on difficult issues in programming the ATRON robot, such as selecting an appropriate role for a module based on the physical configuration of the robot, and executing a sequence of distributed actions.

edit

Project staff

Kasper Støy, Associate professor, Ph.D. [6]
Ulrik Pagh Schultz, Associate professor, Ph.D. [7]
Mirko Bordignon, Ph.D. student
Danish Shaikh, Research Assistant

edit

Collaborators

David Brandt, Ph.D. student
David Christensen, Ph.D. student [8]
Lars Kristian Lindegaard Mikkelsen, M.Sc. student

edit

Alumni

Nicolai Dvinge, M.Sc.

Retrieved from "http://ecosoc.sdu.dk/coe/Morphing_production_lines"