Fraunhofer-Institut für Produktionsanlagen und Konstruktionstechnik IPK
Now showing 1 - 10 of 13
PublicationOn the optimization of automobile panel fitting( 1996)
;Hsieh, L.-H. ;Seliger, G.Qian, W.-H.The difference between the sum of squared deviations and the maximum deviation was studied. The popular least square method refers to the former, but the criterion of many engineering geometrical problems, including fitting problems, often refers to the latter. As a typical case study, this paper presents a new approach to the optimization of the automobile panel fitting, aiming at minimizing the maximum deviation, based on a systematic analysis of the differential adjusting motion's impact on the variations of gaps/flushes. Comparative calculations show: the conventional least square method may lead to a result with the maximum deviation over thirty percent higher than the actual optimal solution, obtained by new method. Moreover, the latter can cope with different tolerances of different seams explicitly.
PublicationSelection of measuring points and optimization algorithm for automotive door fitting( 1996)
;Quian, W.-H. ;Hsieh, L.-H.Seliger, G.The pose uncertainty of the door rleative to the poly-in white robotized assembly can be tackled with a differential adjusting motion of the robot to keep the clearance between them close to the required. We bound that the popular least square method does not denon of such fitting problems refers to the maximum deviation rather than the sum of squared deviations. Aiming at minimizing the former, a new approach to the optimal adjusting during automotive door assembly is presented. Comparative calculations show: the conventional least square mthod may lead to a result with the actual optimal solution, optained by the new method. This paper focuses on the selection of measuring points and the opimization algorithm.
PublicationAutomated learning system for control and supervision of assembly systems( 1995)
;Groth, A. ;Hsieh, L.-H.Seliger, G.Automation has not been applied as widely in assembly as in other manufacturing areas, mainly due to the complexity of operations requiring control of many geometric and technological process parameters. This paper shows that automated learning can be applied to control and supervision of assembly systems for which only a qualitative process model exists. A qualitative process model is built from incomplete a-priori knowledge of processes, facilities and products including essential performance targets. Automated learning algorithms determine optimum process control and supervision strategies based on performance of the assembly system. The process model is updated in real time according to the process results. Critical parameters can be identified and supervision strategies optimized. A prototypical automated bonding system will serve as a practical example showing how automated learning can help determining and verifying control strategies.
PublicationNavigation and control of a mobile climbing robot for shipyards( 1994)
;Seliger, G. ;Hsieh, L.-H.Müller, H.This paper describes the navigation and control features of the mobile climbing robot MoRoMAG which has been developed for welding, gas cutting and other manufacturing processes on shipyards. navigation tasks are solved by dead reckoning navigation and laser beam guidance. For weld seam tracking the robot can be equipped with either a laserscanner, an arc sensor or an inductive sensor. The user friendly programming of the robot facilitates an easy teaching-in of a large number of parameters. By using mobile climbing robots an increase of productivity can be expected.
PublicationA flexible laser scanner system for on-line process monitoring( 1994)
;Seliger, G. ;Schuler, H.Hsieh, L.-H.This paper describes a geometrical sensor system for on-line process monitoring. The sensor consists of a double ligth stripe laser scanner and a modular software for sensor data processing which have been developed in the scope of an EC ESPRIT III project. The paper presents firstly the industrial requirements revealed by a market survey and the overall system architecture. In the second part of the paper the developed novel software algorithms and an example application of the sensor system are described in detail.
PublicationProgramming and navigation of a mobile welding robot( 1994)
;Seliger, G. ;Hsieh, L.-H.Müller, H.Tis paper describes the programming and navigation features of the mobile climbing robot MoRoMAG which has been developed for welding of ship sections and large gas vessels. For seam tracking the robot can be equipped with either a laserscanner, an arc sensor or an inductive sensor. Navigation tasks are solved by dead reckoning navigation and laser beam guidance. The user friendly programming of the robot facilitates the teaching-in of large numer of welding parameters required by automated welding. By using mobile climbing robots an increase of productivity can be expected. The application of MoRoMAG in other manufacturing processes is the subject of current research.
PublicationErkennung formvarianter Geometrien mit einem Laserscanner auf der Basis evolutionärer Algorithmen( 1994)
;Seliger, G. ;Hsieh, L.-H. ;Voit, L.Schuler, H.In this paper, we describe a new geometry measurement system with a robot guided laser scanner which is capable to handle new applications by its recognition and configuration flexiblity. By the use of an optimization algorithm basing on revolutionary algorithms it is able to determine both, geometry and position deused for guidance of automated assembly processes. By the example of sensor guided welding of car door hinges it will be shown how the new geometry recognition system tackies automated robot guided assembly.
PublicationComputer-aided planning system for multiple laser-scanner applications( 1994)
;Schuler, H. ;Hsieh, L.-H.Seliger, G.Laser scanners are efficient tools for on-line inspection of workpiece dimensions and for closed-loop process control.Relevant industrial applications require frequently multiple scanners to acquire many significant dimensions of the workpiece at the same time.One of the most crucial tasks hereby is to determine the spatial arrangement of multiple scanners.Until now, this task has been achieved by triel and error.Heuristic rules are available for simple geometry but not available for complicated workpiece geometry.The effectiveness of this planning approach can only be checked after the physical system installation.This paper presents a computer-aided planning system which derives an optimal spatial arrangement of the multi-scanner setup for 2-dimensional measurement tasks automatically.The necessary inputs of the system are the geometry of the workpiece, the regions of interest on the workpiece and the technical specifications of the sensors.Effects of changing properties such as accu racy or range of the selected sensors can be studied through a realistic sensor simulation. (RSS).The accuracy of the multi scanner setup can be verified on the planning system by specifying the accuracy and the relative arrangement of the single scanners.
PublicationGußgraterkennung mit optischen profilmessenden Sensoren für robotergeführtes Entgraten von Gußteilen( 1991)
;Hsieh, L.-H. ;Seliger, G.Yi, H.-C.Sensoren gewinnen beim Robotereinsatz in modernen Produktionssystemen an Bedeutung, ein Beispiel dafür ist das robotergeführte Entgraten von Gußteilen. Der Grat des zu bearbeitenden Werkstückes muß erst räumlich erfaßt werden, eine Möglichkeit hierzu stellt ein Laser-Scanner dar, der aus den nach dem Triangulationsprinzip berechneten Abstandsinformationen einen Profilschnitt erzeugt. Die erfaßten Profildaten müssen zur Ableitung der geeigneten Werkzeugwege verarbeitet werden, hierfür wurde ein Gußgraterkennungssystem entwickelt. In diesem Beitrag wurde das realisierte System vorgestellt. Die spezifischen Probleme bei der Profilmessung, der Identifizierung eines Gußgrates und der Ermittlung der Gratfußpositionen werden diskutiert. Anschließend wurde ein Anwendungsbeispiel des entwickelten Systems für die Programmierung des Roboters beim Entgraten gezeigt.
PublicationAutomated path generation for sensor-guided deburring( 1989)
;Hsieh, L.-H. ;Lennartz, K.D. ;Yi, H.-C.Seliger, G.This paper discusses the application of sensors to automatically identify the burrs on castings and to plan the robot trajectories for cutting operations. Alternative methods to sensor approach are reviewed and the reasons for using sensors in this task are explained. The prerequisites of an automated path generation are investigated. A realized system using laser-scanner as sensor head and a commercial robot as movement unit is described. Based on this configuration sofware modules for deburring of castings have been developed.