Teleoperation of SCARA with neural network based controller

english

2008

volume 54, issue 2

Univerza v Ljubljani, Fakulteta za strojništvo

This paper describes the development of neural network based controller for the teleoperation of Selective Compliance Assembly Robot Arm (SCARA). The SCARA is controlled and teleoperated via the internet. Presented experiment isused by students at the University of Maribor as a remote educational tool. Application is based on MATLAB/Simulink and LabVIEW software packages. MATLAB/Simulink and developed library DSP-2 Library for Simulink are used for neural network control algorithm development, simulation and code generation. The executable code is downloaded to the Digital Signal Processor (DSP). The DSP controls through the analog and digital I/O the real process and maintain the data connection with the laboratory server. The LabVIEW virtual instrument(VI) is used as a client server application for the teleoperation. LabVIEW VI provides the ability for parameter tuning, signal monitoring, on-line analysis and via Remote Panels technology also teleoperation by using the internet browser (Internet Explorer). The main advantage of a neural network controller is the exploitation of its self-learning capability. For example: when friction or an unexpected disturbance occurs, the user of a remote application does not need any information about the changed robot dynamics, because it is estimated independently of the remote user.

14.07.2011