This paper describes the use of a magnetic levitation haptic device (MLHD) to study the psychophysics of tex- ture roughness. Studies of texture roughness perception per- formed using real textures can be time consuming and ex- pensive. By using a MLHD to simulate texture we are able to quickly and easily adjust texture parameters. A dithered textured surface composed of conical elements is simulated using a constraint surface algorithm. The constraint sur- face shape is defined by the geometry of the elements as well as the size and shape of the virtual probe. The spac- ing of the elements and the size of the probe can be varied continuously and in real time.

Just noticeable difference (JND) experiments were con- ducted over the parameters of probe radius and texture spacing. The JND of roughness was determined with re- spect to element spacing using unforced weighted up-down adaptive threshold estimation. JND?s were found to vary for texture spacing and probe size. JND?s for constant probe size decreased with increasing texture spacing to a mini- mum and then increased again. JND?s for constant spacing increased as probe size increased. These results are con- sistent with a geometric model of probe-texture interaction.