Keywords: nAMT model of human tracking behaviour

ABSTRACT - A nonlinear generalization of the Adaptive Model Theory, nAMT, is compared with human open-loop tracking data across the same range of conditions. The resulting simulations produced effects that mirrored the closed- and open-loop characteristics of the experimental response trajectories. This supports the use of an internal feedback loop for the inversion of external systems in the nAMT model. Other control-systems models (both AMT and feedback-error learning) were unable to reproduce the observed disparity between closed- and open-loop results without fundamental modification. A low internal feedback loop-gain, incorporating a substantial derivative component, caused this effect. This low gain produced acceptable performance due to the relatively low target bandwidth used in the study, allowing the feedback control component to function. Maintenance of the loop-gain at the lowest possible levels is thought to maximize the internal stability of the inverse. The simulation work confirmed that the nAMT model is capable of reproducing human behavior under a wide range of conditions.

----------------------------------------

[...] The open-loop results do not suggest that the human subjects are attempting to reach a transfer function of unity, representing accurate tracking (as has been observed for closed-loop tracking). In fact, a quite different relationship between target and response appears to have emerged.

[...] The nAMT computational model succeeded in generating responses which reproduced many of the principle characteristics of the response trajectories obtained during the experimental study.
[...] In nAMT, using a low internal feedback loop-gain maximizes the stability of the loop but also essentially filters the motor response. This prevents the inverse from becoming accurate at all frequencies.
[...] Paradoxically, the standard AMT model, FEL and, to our knowledge, all other control-systems-type motor control models are incapable of reproducing these results due to the accuracy of the inversion techniques they employ.