AN ARTIFICIAL NEURAL NETWORK METHOD FOR PREDICTING LOWER LIMB JOINT MOMENTS FROM KINEMATIC PARAMETERS DURING COUNTER-MOVEMENT JUMP
Keywords: artificial neural network, joint moment, CMJ
AbstractThe purpose of this study was to develop an artificial neural network (ANN) model for predicting the joint moments of lower limbs using solely the kinematic parameters during counter-movement jump (CMJ). Nine female volleyball players performed CMJ. The joint moments were calculated from experimental data by inverse dynamics (called “measured joint moments” in this study). A “303-3-303 ANN model” was developed with 303 neurons in input layer, three neurons in hidden layer and 303 neurons in output layers. The input variables were the left lower limb extension / flexion joint angles, and the output variables were left lower limb extension / flexion joint moments. The results revealed that the ANN model fitted the experimental data well indicating that the model developed in this study was feasible in the assessment of joint moments for CMJ.
Authors can retain copyright, while granting the International Society of Biomechanics in Sports (ISBS) the right of first publication.