Prediction of 2-D Lift Motions Using Artificial Neural Networks: An Evolutionary Approach

Movement prediction is an important aspect of human simulation, where more efficient and accurate models are needed. Artificial neural networks could potentially serve as a modeling option in this realm. This investigation evaluates the performance of a particular artificial neural network structure in modeling sagittally symmetric two-dimensional lifting and lowering movements. Model performance was evaluated using three training datasets, each consisting of distinct representation levels of the overall dataset. Results are discussed in terms of their practical meaning, and suggestions for future improvements in the modeling scheme are provided. Overall, artificial neural networks show promise as a modeling paradigm for the prediction of human movement.