Human force proprioception has previously been shown to be important in manual control tasks. Recent experiments suggest that force proprioception is more accurate than position proprioception. Still, knowledge about proprioceptive qualities and their relevance in manual control is lacking or inconsistent. This research aims to experimentally establish the respective difference in force and position perception performance and attempts to demonstrate its influence on the neuromuscular system in manual control tasks and on which manipulator properties are optimal. The first experiment confirmed that the force sensing Golgi Tendon Organ (GTO) is more accurate than position sensing Muscle Spindles, by measuring smaller human just noticeable differences of force than of position in an experiment of discriminating side-stick manipulator stiffness at various conditions.
Offline simulations of a detailed model of the neuromuscular system in a manual control task validated higher GTO activity with non-zero stiffness manipulators than zero stiffness. The second experiment measured human control behavior in a double integrator controlled element pursuit task with varying manipulator stiffness intended to affect GTO activity. Zero stiffness manipulators showed worse tracking performance and lower degree of linearity of human control inputs which is related to less accurate control actions and higher uncertainty of quasi-linear human operator models. The two experiments combined demonstrate that humans' neuromuscular force sensors are more accurate than position sensors and that higher manipulator stiffness inducing higher force sensor activity results in more accurate control behavior.