The tactile display of vector data
The number of available perceptual dimensions upon which to encode complex scientific information on a computer screen is restricted. This thesis presents what is believed to be the first study of the use of tactile sensations for scientific information display. Perceptual aspects of scientific visualisation and sonification (the use of sound to portray information) have been reviewed in parallel with the perceptual capabilities of the tactile sense. Current and emergent tactile display technologies, and current uses of such devices have also been reviewed. Vector information was chosen to be displayed as it was concluded that it was difficult to display visually and aurally, but well suited to tactile display. This, and other conclusions led to the design of a novel tactile display system. The display comprises a 3x3 grid of electrocutaneous stimulators on the fingers of the non-dominant hand. An extension of the animation of the phantom sensation phenomena into two dimensions was created to encode a vector. Thus a wave of stimulation, whose speed and direction are determined by the vector information, is felt to move over the fingers of the hand. Observers found the display of the vector information to be intuitively interpretable. The stimulations generated by the discrete stimulators were fused into one, well-defined sensation. A series of experiments were conducted to evaluate the system: To determine if the encoding schemes chosen to portray vector magnitude and direction were orthogonal, and to generate psychophysical mapping functions to allow vector information from a scientific data set to be encoded in a perceptually linear manner. The encoding of vector information, magnitude onto the speed of the tactile wave, and direction onto the direction of the wave, was found to show a high degree of orthogonality. The quantitative results show that psychophysical mappings would be simple, a power law in the case of the magnitude/speed, and linear in the case of direction/direction. Vector acuity is difficult to express numerically, as it is affected by the choice of variance criteria and experimental methodology but is of the order of 3-1 0° for direction and one part in 35 to one part in 15 for magnitude. These results are considered good and show great promise for the device. The mean decision time was 13 seconds, however, this is considered to be an artefact of the experimental design. Further experimental work is proposed to explore other encoding schemes, and obtain a true measurement of the time taken to perceive the stimulus. The study of the device when integrated into a scientific perceptualisation system is also explored.