This paper investigates the practical application of neural networks with online learning for robot manipulator control. Experiments have been performed with a PUMA Unimate 560 industrial manipulator with reference signals of increasing complexity. The results presented demonstrate features of the proposed framework such as accurate trajectory tracking, online learning, and bounded weights, which guarantee closed loop stability.
|Publication status||Published - 2006|
|Event||International Control Conference (ICC2006) - Glasgow, United Kingdom|
Duration: 30 Aug 2006 → 1 Sep 2006
|Conference||International Control Conference (ICC2006)|
|Period||30/08/06 → 1/09/06|