Linearity of human commands for telerobotics

We identify frequency limitations on linearity in two-dimensional (2-D) motion for three subject groups: young (mean and σ: 23±3), old (72±2), and movement-disabled (65±16). The linearity of the relationship between input and output-i.e., between target tracking signals and resulting human upper extremity motion-is quantified using adaptive coherence estimation. We show that linearity decreases with input frequency. Old and disabled subjects´ 2-D motion is almost completely nonlinear. Nonlinear filtering therefore will likely improve precision in rehabilitative and surgical teleoperation