Objective. To develop kinematic structure models of the classical moves for highly skilled skiers using the computer neural networks, and to improve their technical actions on this basis.
Methods. Analysis of scientific and methodical literature; analysis of athlete’s medical record; video shooting; biomechanical video-computer analysis and neural networks modeling of the movement technique of highly skilled hearing-impaired skiers; pedagogical experiment; mathematical statistics.
Results. The modeling kinematic schemes of the diagonal stride and double poling technique of highly skilled skiers with hearing impairments have been made by means of “Biovideo” software, kinematic indices of the technique have been identified. The neural networks of multilayer perceptron type have been developed as a simulation of the velocity of the skier’s general center of mass in the movement cycle of diagonal stride and of radial-basis function as double poling classical moves. The technology of improving technical actions of highly skilled skiers with hearing impairment while using classical technique in the annual training cycle has been developed, as a result of using which during annual preparation macrocycles of the Deaflimpic team of Ukraine in ski races a statistically significant improvement (p <0.05) of technique kinematic indices has occurred.
Conclusions. Use of neural network modeling in the process of technique improvement of highly skilled skiers has been substantiated.