Electrogoniometer error compensation, for hand position measurement during planar reaching movements

Abstract

Reliable and valid measurements of the shoulder and elbow angles and the hand position is an essential requirement for assessment of the motor impairments and improvements in upper limb rehabilitation systems such as robot therapy. Flexible electrogoniometers are used in various clinical applications to record body joint angles. Electrogoniometers are susceptible to some errors due to the rotation between two endblocks, the physical characteristics, and the effect of the body. This study aimed to evaluate and compensate for the errors of the shoulder and elbow electrogoniometers simultaneously in planar reaching movements and to propose a method to measure the correct hand position using the electrogoniometers. First, the calibration of the electrogoniometers and compensation for the body error was performed using polynomial mappings separately. Then an artificial neural network (ANN) was used to map the calculated position trajectories to the desired trajectories. Despite using one neural network for four different planar movements, the results showed an accurate matching between the ANN corrected and desired trajectories with average RMSE < 2.5 cm and average VAF > 97%. The results of this study can be used for reliable measurement of the hand position using electrogoniometer data in movement rehabilitation systems.