Home // International Journal On Advances in Intelligent Systems, volume 7, numbers 3 and 4, 2014 // View article

Enhancing Robustness through Mechanical Cognitivization

Authors:
Gideon Avigad
Avi Weiss
Wei Li

Keywords: Cognitive robotics, developmental robotics, evolutionary algorithms

Abstract:
The common approach for training robots is to expose them to different environmental scenarios, training their controllers to have the best possible commands when untrained scenarios are encountered. When humans train they do the same. They try new manipulations by performing within different environments. However, humans training (and in fact development from infancy to maturity) also includes a type of training which, although claimed to improve cognitive capabilities, has not, to date, been adopted for the training of robots. This type of training involves the restriction of manipulation capabilities while performing different tasks, e.g., climbing with just one hand. Recently a research that facilitates functions instead of a mechanical systems that aims at exploring the invigorating idea that such training, would enhance the robustness of robots, has been published. This type of training has been termed as Mechanical Cognitivization. In the current paper, the preliminary published results are detailed and more elaborated examples are given. Specifically, it is shown that the Mechanical Cognitivization based training improves the performances when performing within untrained environments and when malfunctions occur. The advantages of the suggested training are highlighted through facilitating a comparison between two schemes that include a common neural net (with no training of restricted modes) and the recently introduced Mechanical Cognitivization based neural net for which the training includes training of restricted modes. The results highlight the advantages of Mechanical Cognitivization based training in enhancing robustness.

Pages: 652 to 661

Copyright: Copyright (c) to authors, 2014. Used with permission.

Publication date: December 30, 2014

Published in: journal

ISSN: 1942-2679