Rowing stroke force estimation with EMG signals using artificial neural networks

Performance analysis in sports activities such as rowing requires measurement of athlete hand force. The use of inexpensive and easily portable active electromyogram (EMG) electrodes and position sensors would be advantageous compared to the use of heavy duty expensive force sensors that require bulky frames and are vulnerable to overload. In this study, artificial neural networks (ANN) are employed for hand force estimation using EMG signals collected from upper arm muscles involved in elbow joint movement and sensed elbow angular position and velocity. In particular, the performance of multilayer perceptron (MLPANN) and radial basis function ANN (RBFANN) for hand force estimation under emulated rowing condition are compared experimentally