An Evaluation of Hand-Force Prediction Using Artificial Neural-Network Regression Models of Surface EMG Signals for Handwear Devices

Table 5

A Comparison of prediction methods of handgrip forces in intrasession evaluations.

Regression method	Position (# muscles)	Type of electrode	Max. force	Measurement	Errors
Linear [15]	Forearm (6)	Paste	300 N	Static	52–80 N (MAE)	—
				Dynamic	27–41 N (MAE)	—
L1-ANN [14]	Forearm/Hand (3)	Paste	50 N	Dynamic	—	0.77
L4-ANN (Ours)	Hand (4)	Dry	200 N	Static	45 N (RMSE)	0.85