Naturally Supervised Learning in Manipulable Technologies

Abstract

The relationship between physiological systems and modern electromechanical technologies is fast becoming intimate with high degrees of complex interaction. It can be argued that muscular function, limb movements, and touch perception serve supervisory functions for movement control in motion and touch-based (e.g. manipulable) devices/interfaces and human-machine interfaces in general. To get at this hypothesis requires the use of novel techniques and analyses which demonstrate the multifaceted and regulatory role of adaptive physiological processes in these interactions. Neuromechanics is an approach that unifies the role of physiological function, motor performance, and environmental effects in determining human performance. A neuromechanical perspective will be used to explain the effect of environmental fluctuations on supervisory mechanisms, which leads to adaptive physiological responses. Three experiments are presented using two different types of virtual environment that allowed for selective switching between two sets of environmental forces. This switching was done in various ways to maximize the variety of results. Electromyography (EMG) and kinematic information contributed to the development of human performance-related measures. Both descriptive and specialized analyses were conducted: peak amplitude analysis, loop trace analysis, and the analysis of unmatched muscle power. Results presented here provide a window into performance under a range of conditions. These analyses also demonstrated myriad consequences for force-related fluctuations on dynamic physiological regulation. The findings presented here could be applied to the dynamic control of touch-based and movement-sensitive human-machine systems. In particular, the design of systems such as human-robotic systems, touch screen devices, and rehabilitative technologies could benefit from this research.