An identification of body segment parameters in the upper extremity

This paper presents a method for identifying static parameters of all three upper extremity segments. An industrial robot was used for moving the subject's arm along a specified sagittal trajectory during which measurements of the shoulder, elbow and wrist angles and forces in the contact point were collected. An identification procedure was developed in analogy with those performed on industrial robots, which assumes the upper extremity model of a 3DOF planar structure in a closed kinematic chain. At first, a suitable trajectory had to be obtained in all three joints with appropriate low speed and with very small angle deviations around an operating point. The arm was assumed to be linear in this point since passive properties of the arm showed no nonlinearities for such small angular deviations. Since the imposed movements were very slow all dynamic effects could be discarded, which simplified the mathematical complexity. By programming a suitable trajectory into the robot controller the ill conditioning problem could be improved significantly. The algorithm eventually identifies masses and mass center coordinates for the lower arm and palm segments, their products for the upper arm and the operating point passive moments (summed elasticities and Coulomb frictions) for all three joints in the sagittal plane. The results are eventually compared to the literature estimations which are based on average population.