Affiliations: Department of Mechanics, Faculty of Mechanical Engineering, CTU in Prague, Technická 4, 16607 Prague 6, Czech Republic | [x] Mechanical Engineering Faculty, Laboratory for Materials Technology, University of Applied Science, Regensburg, Germany | [y] University Clinic, Department of Traumatology, Regensburg, Germany
Abstract: Today artificial neural networks can be trained to solve problems that are difficult for conventional computers or human beings. The big advantage of an artificial neural network is results obtained without knowledge of the algorithm procedure or without full and exact information. Therefore an artificial neural network was used to predict the muscle forces. The aim of the study was to simplify prediction of muscle forces which are difficult to determine, because many muscles act cooperatively. However, orthopeadists, biomechanical engineers and physical therapists need to take muscle forces into consideration because joint contact forces, as well as muscle forces, need to be estimated in order to understand the joint and bone loading. In terms of sensitivity of the muscle parameters to the results from the proposed neural network object, the muscle force prediction was simplified.
