Carnegie Mellon Robotics Institute
doctoral dissertation, tech. report CMU-RI-TR-08-45, Robotics Institute, Carnegie Mellon University, August, 2008
|This thesis describes the creation of an experimental platform for the exploration of human texture
perception and its relationship to physical phenomena. Virtual textures are simulated and their
physical and psychophysical properties are determined.
Experiments were conducted into the ways in which modeling both texture and probe geometry influence the perception of virtual surfaces. The effects of compliance and friction on texture perception were also explored. Roughness detection thresholds were documented and their relationship to probe geometry and compliance were examined. A spectral analysis of the force signal demonstrated that force variability characterized roughness and that FA1 neural receptors were the receptors primarily responsible for mediating indirect roughness perception.
Psychophysical studies of texture perception have been hampered by the need for the expensive manufacture of finely textured surfaces. These texture samples provide only discrete levels of experimental variables for study and are time-consuming to present to subjects. Virtual haptic textures provide continuous control of variables and are quick and easy to use. Unfortunately, psychophysical experimental results seem to differ between real and virtual textures.
An algorithm was used which modeled probe and texture geometry to render virtual haptic texture that was perceptually the same as real texture. The texture was created by constraining the haptic device manipulandum to follow a surface dictated by the interaction of the size and shape of the virtual probe and texture elements. Compliance and friction effects were also simulated.
Subject roughness magnitude estimations were used to determine the psychophysical function for roughness. The effect of probe geometry, compliance, and friction on this function and on roughness thresholds was determined. Force and position recordings were correlated with roughness estimates and used to evaluate the physical properties characterizing roughness perception.
This thesis provides significant new insight into the perception of roughness and its relationship to texture geometry, compliance and friction. It also establishes an experimental platform and framework which should expedite future studies of texture perception.
Associated Lab(s) / Group(s):
Microdynamic Systems Laboratory
Number of pages: 298
|Bertram Unger, "Psychophysics of Virtual Texture Perception," doctoral dissertation, tech. report CMU-RI-TR-08-45, Robotics Institute, Carnegie Mellon University, August, 2008|
author = "Bertram Unger",
title = "Psychophysics of Virtual Texture Perception",
booktitle = "",
school = "Robotics Institute, Carnegie Mellon University",
month = "August",
year = "2008",
address= "Pittsburgh, PA",
|The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.|
Contact Us | Update Instructions