Stimulus identification in the excitation of the corneal nerve in an aviar model by neural networks

Abstract

This paper describes the application of differential neural networks (DNN) to classify the inverse response of the optical nerve in an aviar model. Generally speaking, the main objective in signal classification is to obtain the class of an obtained response given a characteristic stimulus. This work deals with the inverse procedure, that is, to recover the stimulus with the measure of the optical nerve response after a classification. A robust exact differentiator (RED) based on the super-twisting algorithm (STA), that is a second order sliding mode technique, reinforces the DNN classifier. The main idea is to combine the DNN and RED approaches to implement a so-called nonlinear observer for unknown inputs. The results show how the input stimulus in the eye of an aviar model are reproduced with the DNN-STA scheme based on a previous classification obtained measuring the response in the optical nerve with the unknown stimulus.