Fractional Order Adaptive Control for Manipulator Systems


Bencsik A. L.

9th WSEAS International Conference on Simulation, Modelling and Optimization, Budapest, Macaristan, 3 - 05 Eylül 2009, ss.464-465 identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: Budapest
  • Basıldığı Ülke: Macaristan
  • Sayfa Sayıları: ss.464-465
  • İstanbul Teknik Üniversitesi Adresli: Hayır

Özet

The relatively slow progress in the subject of sensor-based robot intelligence and autonomy naturally made the interest in the development of haptic devices and master-slave systems increase. Such systems can roughly be categorized as devices of high precision and limited work-load, and instruments developed for dealing with great work-load and limited precision in harsh industrial environment like in the case of deburring iron casts. In this latter class hydraulic actuation shows crucial advantages in comparison with electric and pneumatic systems. Aiming at this potential field of application in the labs of Budapest Tech a 6 Degree Of Freedom (DOF) hydraulic joystick-like master arm has been developed in which reflecting and sensing the necessary forces is solved by a hydraulic subsystem, too. In the first part of this paper a brief overview of the master-slave development is presented. In the rest potential improvement of the already achieved results is summarized on the basis of a detailed mathematical model of the electric servo valve controlled hydraulic differential cylinder with especial emphasis on the adhesion between the cylinder and the packing ring of the piston. Adhesion is roughly nonlinear effect causing abrupt change in the direction of the friction forces in the zero-transmission of the piston-cylinder relative velocity. Traditional control applying PID-type feedback yields poor results near zero velocity where the behavior of the friction forces is roughly nonlinear, and almost singular. The operation of a proposed adaptive control using fractional order derivatives is discussed and demonstrated by simulation results in the second part of the paper.