Dynamic fuzzy neural network for the intelligent control of a humanoid robot

作者：Tang Zhe^*; Er Meng Joo; Qi Fang

来源：Computer Systems Science and Engineering, 2011, 26(3): 187-196.

Intelligent technique is a useful tool for complex system modelling, where there is a large amount of training data instead of a clear mathematical model. The handling of megadata is one of key issues in the field of e-science. In this paper, an efficient intelligent technique, DFNN (Dynamic Fuzzy Neural Network), has been implemented for humanoid robot modeling. By this implementation, it is well demonstrated that the intelligent technique can be used derive some control disciplines from large data volumes. The inputs of the DFNN are two key parameters which represent a specific walking gait, while the output is an evaluation value of humanoid walking performance. Based on the evaluation information, the middle level control can select optimal walking gaits to realize a global path commanded from the high level. Simulation results are given to demonstrate the effectiveness of the intelligent control method.

• 出版日期2011-5
• 单位中南大学