استفاده از برنامه نویسی ژنتیک برای درجه بندی روبات های صنعتی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|18480||2007||10 صفحه PDF||سفارش دهید||6104 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Computers in Industry, Volume 58, Issue 3, April 2007, Pages 255–264
Robot calibration is a widely studied area for which a variety of solutions have been generated. Most of the methods proposed address the calibration problem by establishing a model structure followed by indirect, often ill-conditioned numeric parameter identification. This paper introduces a new inverse static kinematic calibration technique based on genetic programming, which is used to establish and identify model structure and parameters. The technique has the potential to identify the true calibration model avoiding the problems of conventional methods. The fundamentals of this approach are described and experimental results provided.
The range of applications of industrial robots has expanded in recent years in part due to advances in programming capabilities resulting from the development of offline-programming (OLP) systems , ,  and , which enable program development to take place in a virtual environment. However, the critical factor in the application of OLP systems is the accuracy with which they are able to model the physical robot. Deviations between the idealized simulation model and the real world cause the OLP system to generate robot poses with large positional errors. Any successful off-line programming procedure must therefore include a method of compensating for the errors between the simulation and the actual robot. Robot calibration techniques ,  and  are designed to improve the software model of the robot so that it is able to more closely represent the behaviour of the actual robot. These techniques can be classified into two types either static or dynamic calibration. This paper addresses the problem of static  calibration, more specifically inverse static calibration. The aim is to improve the kinematic model used to relate the position of the robot end-effector with the joint sensor readings so that it more accurately represents its actual position.
نتیجه گیری انگلیسی
This paper demonstrates the potential of the symbolic calibration method based on evolved joint correction models. The experimental results using a three joint correction model show a significant reduction in tool point error. The advantage of the method, which has been implemented as distal supervised learning is the automatic generation of the correction models using genetic programming, i.e. symbolic regression. Classical calibration methods require human involvement to establish a calibration model, which is subsequently fitted to calibration data employing methods from numerical analysis. The method used in this work combines the processes of automatically generating and evaluating correction models in a direct evolutionary search algorithm. Applied to a complete and accurate kinematic model  the method has the potential to solve the calibration problem by finding suitable or even true parameter values. In addition, since it is not confined to a fixed model structure the method is more flexible than classical numerical calibration techniques.