Skip to main content
European Commission logo print header

Skill-based Inspection and Assembly for Reconfigurable Automation Systems

Final Report Summary - SIARAS (Skill-based Inspection and Assembly for Reconfigurable Automation Systems)

The aim of the SIARAS project is the knowledge based automatic reconfiguration of automation systems. Therefore modules were implemented on the one hand for the representation of skills and parameters of devices and on the other hand for representation of the process flow and bounding conditions. For finding suitable configuration, a skill server reasoning framework with search and optimisation algorithms was developed to match the process description with the skill representation.

A given task description is needed that describes the principle steps of the task and needed requirements, boundary conditions or tolerances. Therefore, the SIARAS consortium developed a flow chart-based approach, were the user can describe and parameterise the process description step by step.

With the task description, the skill server is able to compare the needed requirements in the task description with his knowledge base. Therefore, search strategies were developed and implemented to perform a comprehensive search in the knowledge base. The user can limit the search by using quality criteria (like cost or accuracy) or he can chose the devices himself by all skill server suggestions.

With the skill server reasoning framework, an automated selection and parameterisation of devices is possible to solve reconfiguration task. The consortium provides a complete framework that supports the user with an interactive and graphical reconfiguration process.

It is very important for the user, that he can recognise the reconfiguration process. Thus, the SIARAS consortium developed a structured method for in interactive reconfiguration process. That means, the user changes the task description and the skill server shows the user step by step the need for reconfiguration and possible suggestions.

To allow the skill server to reason on questions concerning the work piece, the information about it and the data model has to be accessible to the skill server. All data formats for exchange of CAx data have one main disadvantage: they are not accessible in the sense to add some metadata. Some of the data formats offer at least the possibility to use different profiles. However, as a consequence, new parsers and tools had to be written, as the available tools do not support new profiles. Therefore, the requirements for the SIARAS skill server workpiece representation were specified independently of the available data formats.

With this data model, the SIARAS consortium created a workpiece model that can be used for automatic reasoning. All relevant information can be stored as meta-data in current workpiece description of CAx data. The skill server can read this information and use it in the reasoning process for finding suitable devices that can handle the workpiece properties.

The framework for the integration of simulation tools in the skill server concept is a client-server solution that provides a general interface for formulating and asking simulation queries (here called query interface). The interface exposes a black box simulation model (here called query model) towards the skill server with only search parameters visible. A query model together with a set of parameters and a chosen question forms a query which can be asked through the query interface.

In the field of robotics some tools exist that can be used for local parameter calculation, e.g. trajectory planer for robot arms. However for global parameters like cycle times, energy consumption or the position of the devices no optimisation tools are available. Thus, during the project the consortium developed algorithms and interfaces for scanning the parameter space and optimise the process with the help of simulation tools.

With the implemented interface and algorithm, the skill server can optimise chosen configuration in a global way. Global parameters like cycle time or energy consumption can be improved. The principle was realised with the genetic algorithm toolbox of Matlab and simulation tool 3DCreate.

ABB has developed a tool called Robot Studio Deployment Wizard. This tool is used to react on reconfiguration suggestions from the skill server. The tool will make use of the skill server to provide setup support for process and equipment work cell changes. The deployment wizard will then identify these changes, generate the needed cell setup and help the user through all the steps necessary to calibrate the work object, frames, and tools used in the cell.

This is accomplished by analysing the models, finding the reference points to be used for calibration and selecting the calibration method. This can be linked with a common control tool to make the reconfiguration process easier. In addition, three-dimensional (3D) snapshots are taken from the offline simulation tool, and combined with a teach pendant application program which presents the images on the teach pendant and guides the user step-by-step through the calibration process. This is combined with a robot program which can be used to quickly move the robot to all of the entered points (for calibration checks) and to save the points for future verification.