Análisis gráficode la neuro-dinámica en un robot capaz de realizar aprendizaje imitativo Graphic analisys ofthe neural-dynamics in a robot able to do imitative learning

Multilayer Perceptron is an artificial neural network able to learn to work in a specific desired way through examples, and once it is trained, it is able to react correctly in situations for which it was not specifically trained, due to its generalization capability. These examples are specified by the designer or can be assimilated with the support of a self-organized map. This paper shows the training verification and the working analysis of the multilayer perceptron used during the development of an imitation learning model; which is focused to a mobile robot displacement over two-dimensional space. This was done designing a comparison method between the attractor’s graphs generated by the instructor robot and the apprentice robot. This method is supported in digital