Nonparametric identification of wrestlers' eye movements using a differential neural network

This article presents the results of applying oculography to the task of testing martial arts athletes using a virtual environment demonstration in the HTC Vive Pro Eye helmet. Oculographs integrated into a VR-helmet often have a low sampling rate, and are also characterized by possible pupil loss during recording. To combat these effects, it is possible to apply filtering, including the Kalman filter and analogues. In this case, an adequate mathematical model in the state space is required. Usually, a parametric or nonparametric model of the system in question is created. It is not always possible to give an adequate mathematical description of the processes occurring in the system, or the system itself can be represented as a "black box". The oculomotor system may also be referred to such systems. In such cases, nonparametric identification is applicable, that is, identification of the dynamics of the system. In this paper, it is proposed to identify the dynamics of the sys-tem using differential neural networks. The standard sigmoidal activation function was replaced by the Izhikevich activation function described by differential equations. The result of the work of the neural network identifier was an approximate system describing the dynamics of the movement of the eye. A computational modeling was carried out. The workability of the model is investigated on several data sets obtained by recording the oculomotor reaction of athletes-wrestlers to visual stimuli in a virtual environment, the effectiveness of the neural network learning laws is shown.