Objectives and content
Today the high cost of robotic system development and the lack of knowledge of the practical implementation of 'skilled' and adaptive systems have been major limiting factors for the automation of a large variety of assembly and disassembly tasks. This is especially true for tasks that require the ability to handle non-rigid materials like textile, food, plastic foam or wire. This is a great limitation since the field of production tasks where non-rigid materials are handled is extremely wide.
The aim of SKILL-MART is to find generic answers to a number of fundamental questions like: "What is skill?", "Can skill be defined or measured or described in a structured manner?", "Can Programming by Demonstration be used to teach skill?", "Can non-rigid product behaviour be interpreted in terms of skilled manipulation?".
These questions are at a most fundamental level of understanding, but are not without great significance to industry in the medium to long term. The industrial partners require paradigms to be established that will shape their approach to unstructured skilled tasks. The industrial involvement within the project also ensures that the hypothetical can be grounded in the real there is no shortage of real skilled tasks that will prove or disprove the developed paradigms. This close interaction with research teams working on real skilled tasks will lead to solutions to specific tasks which in themselves will act as a foundation to approach the wider questions. To act as the real foundation of the studies a number of industrial tasks have been identified, tasks which are beyond current automation such as forming and installing complex three dimensional wire looms in aircraft or automobiles or packaging flexible products. The research and development taking place in order to automate the real example processes will lead to the creation of a guide containing guidelines and examples for the development and design of learning skilled multi-arm systems, capable of handling non-rigid materials. This guide will meet the industrial objectives of firstly establishing a better understanding and presenting a formal description of skilled manipulation tasks with non-rigid materials. Secondly, it will enable industries to automate processes involving non-rigid materials by applying the knowledge gained through the project.
The research approach followed to achieve these goals starts with analysing a class of tasks concerned with the manipulation of non-rigid materials and goes on by using the analysis results for technical and methodological development on different key areas. Thereafter, the individual results are brought together and integrated resulting in a demonstrator for a chosen task with which testing and validation of the projects outcome is performed.
Funding SchemeCSC - Cost-sharing contracts
BS35 7QW Bristol
29071 Estepona (Malaga)
BS8 1TR Bristol