Bouchiba F.,Laboratory of Power Electronic |
Nouibat W.,Laboratory of Power Electronic
International Review of Automatic Control | Year: 2015
This paper focuses on two fundamental aspects, the development of a learning algorithm of Neuro-Fuzzy Networks (NFN) called the Super Self Adapting Back-propagation with Adaptive Momentum (SSABAM) and its application to the navigation of the mobile robot in an unknown environment. This algorithm is compared to other algorithms such as the Back Propagation (BP), the Back Propagation with Adaptive Momentum (BPAM) and the Super Self- Adapting Back-propagation (SSAB). The results demonstrate that the proposed algorithm, which adjusts automatically the learning rate and the momentum parameter, converges faster than the other algorithms. The optimized NFN controllers by these algorithms allow the robot’s navigation; the quality of the different navigations is evaluated by using the proposed fuzzy expert. Based on the simulation results, after the navigation, the optimized controller by the proposed algorithm is more efficient than the other optimized controllers, in terms of both the time navigation and the fuzzy expert's decision. Finally, the optimized controller by the proposed algorithm is successfully validated by experimental implementation with a real mobile robot. © 2015 Praise Worthy Prize S.r.l. - All rights reserved.