Navigating without vision: basic and applied research.

We describe some of the results of our program of basic and applied research on navigating without vision. One basic research topic that we have studied extensively is path integration, a form of navigation in which perceived self-motion is integrated over time to obtain an estimate of current position and orientation. In experiments on pathway completion, one test of path integration ability, we have found that subjects who are passively guided over the outbound path without vision exhibit significant errors when attempting to return to the origin but are nevertheless sensitive to turns and segment lengths in the stimulus path. We have also found no major differences in path integration ability among blind and sighted populations. A model we have developed that attributes errors in path integration to errors in encoding the stimulus path is a good beginning toward understanding path integration performance. In other research on path integration, in which optic flow information was manipulated in addition to the proprioceptive and vestibular information of nonvisual locomotion, we have found that optic flow is a weak input to the path integration process. In other basic research, our studies of auditory distance perception in outdoor environments show systematic underestimation of sound source distance. Our applied research has been concerned with developing and evaluating a navigation system for the visually impaired that uses three recent technologies: the Global Positioning System, Geographic Information Systems, and virtual acoustics. Our work shows that there is considerable promise of these three technologies in allowing visually impaired individuals to navigate and learn about unfamiliar environments without the assistance of human guides.

Active control of locomotion facilitates nonvisual navigation.

In some navigation tasks, participants are more accurate if they view the environment beforehand. To characterize the benefits associated with visual previews, 32 blindfolded participants were guided along simple paths and asked to walk unassisted to a specified destination (e.g., the origin). Paths were completed without vision, with or without a visual preview of the environment. Previews did not necessarily improve nonvisual navigation. When previewed landmarks stood near the origin or at off-path locations, they provided little benefit; by contrast, when they specified intermediate destinations (thereby increasing the degree of active control), performance was greatly enhanced. The results suggest that the benefit of a visual preview stems from the information it supplies for actively controlled locomotion. Accuracy in reaching the final destination, however, is strongly contingent upon the destination's location during the preview.

Updating of locations during whole-body rotations in patients with hemispatial neglect.

Posterior parietal cortex lesions have been associated with both hemispatial neglect and spatial-updating deficits. Currently, the relation between these processes remains poorly understood. We tested the ability of parietal patients with neglect to update remembered target locations during passive whole-body rotations. The rotations and manual pointing responses were executed with and without vision. During the rotation, the remembered location stayed on the same side of the body midline or crossed the midline. Parietal patients generally underestimated rotations, as compared with control groups, but updated targets equally well on either side of the body midline, regardless of the amount of updating required. Once parietal patients have localized a target, they can use self-motion information to update its location, even if it passes into the region they typically neglect. This lack of contralesional updating effects contrasts with impairments in eye position updating found in previous work with parietal patients.

Is the anisotropy of perceived 3-D shape invariant across scale?

A number of studies have resulted in the finding of a 3-D perceptual anisotropy, whereby spatial intervals oriented in depth are perceived to be smaller than physically equal intervals in the frontoparallel plane. In this experiment, we examined whether this anisotropy is scale invariant. The stimuli were L shapes created by two rods placed flat on a level grassy field, with one rod defining a frontoparallel interval, and the other, a depth interval. Observers monocularly and binocularly viewed L shapes at two scales such that they were projectively equivalent under monocular viewing. Observers judged the aspect ratio (depth/width) of each shape. Judged aspect ratio indicated a perceptual anisotropy that was invariant with scale for monocular viewing, but not for binocular viewing. When perspective is kept constant, monocular viewing results in perceptual anisotropy that is invariant across these two scales and presumably across still larger scales. This scale invariance indicates that the perception of shape under these conditions is determined independently of the perception of size.

Immersive virtual environment technology as a basic research tool in psychology.

Immersive virtual environment (IVE) technology has great promise as a tool for basic experimental research in psychology. IVE technology gives participants the experience of being surrounded by the computer-synthesized environment. We begin with a discussion of the various devices needed to implement immersive virtual environments, including object manipulation and social interaction. We review the benefits and drawbacks associated with virtual environment technology, in comparison with more conventional ways of doing basic experimental research. We then consider a variety of examples of research using IVE technology in the areas of perception, spatial cognition, and social interaction.

Reproduction of object shape is more accurate without the continued availability of visual information.

An unfamiliar configuration lying in depth and viewed from a distance is typically seen as foreshortened. The hypothesis motivating this research was that a change in an observer's viewpoint even when the configuration is no longer visible induces an imaginal updating of the internal representation and thus reduces the degree of foreshortening. In experiment 1, observers attempted to reproduce configurations defined by three small glowing balls on a table 2 m distant under conditions of darkness following 'viewpoint change' instructions. In one condition, observers reproduced the continuously visible configuration using three other glowing balls on a nearer table while imagining standing at the distant table. In the other condition, observers viewed the configuration, it was then removed, and they walked in darkness to the far table and reproduced the configuration. Even though the observers received no additional information about the stimulus configuration in walking to the table, they were more accurate (less foreshortening) than in the other condition. In experiment 2, observers reproduced distant configurations on a nearer table more accurately when doing so from memory than when doing so while viewing the distant stimulus configuration. In experiment 3, observers performed both the real and imagined perspective change after memorizing the remote configuration. The results of the three experiments indicate that the continued visual presence of the target configuration impedes imaginary perspective-change performance and that an actual change in viewpoint does not increase reproduction accuracy substantially over that obtained with an imagined change in viewpoint.

Assessing auditory distance perception using perceptually directed action.

Three experiments investigated auditory distance perception under natural listening conditions in a large open field. Targets varied in egocentric distance from 3 to 16 m. By presenting visual targets at these same locations on other trials, we were able to compare visual and auditory distance perception under similar circumstances. In some experimental conditions, observers made verbal reports of target distance. In others, observers viewed or listened to the target and then, without further perceptual information about the target, attempted to face the target, walk directly to it, or walk along a two-segment indirect path to it. The primary results were these. First, the verbal and walking responses were largely concordant, with the walking responses exhibiting less between-observer variability. Second, different motoric responses provided consistent estimates of the perceived target locations and, therefore, of the initially perceived distances. Third, under circumstances for which visual targets were perceived more or less correctly in distance using the more precise walking response, auditory targets were generally perceived with considerable systematic error. In particular, the perceived locations of the auditory targets varied only about half as much in distance as did the physical targets; in addition, there was a tendency to underestimate target distance, except for the closest targets.

Visually perceived location is an invariant in the control of action.

We provide experimental evidence that perceived location is an invariant in the control of action, by showing that different actions are directed toward a single visually specified location in space (corresponding to the putative perceived location) and that this single location, although specified by a fixed physical target, varies with the availability of information about the distance of that target. Observers in two conditions varying in the availability of egocentric distance cues viewed targets at 1.5, 3.1, or 6.0 m and then attempted to walk to the target with eyes closed using one of three paths; the path was not specified until after vision was occluded. The observers stopped at about the same location regardless of the path taken, providing evidence that action was being controlled by some invariant, ostensibly visually perceived location. That it was indeed perceived location was indicated by the manipulation of information about target distance--the trajectories in the full-cues condition converged near the physical target locations, whereas those in the reduced-cues condition converged at locations consistent with the usual perceptual errors found when distance cues are impoverished.

Comparison of two indicators of perceived egocentric distance under full-cue and reduced-cue conditions.

It has not been established that walking without vision to previewed targets is indeed controlled by perceived distance. To this end, we compared walking and verbal report as distance indicators, looking for a tight covariation in responses that would indicate control by a common variable. Targets from 79-500 cm away were presented under dark and well-lit conditions. Both verbal reports and walking indicated overestimation of near targets and underestimation of far targets under dark viewing conditions. Moreover, the finding that verbally reported distance plotted essentially as a single-valued function of walked distance and vice versa is evidence that both indicators were responding to the same internal variable, ostensibly perceived distance. In addition, binocular parallax, absolute motion parallax, and angular elevation were evaluated as distance cues, and only angular elevation exerted a large influence on perceived distance.

Visual perception of egocentric distance as assessed by triangulation.

Two triangulation methods for measuring perceived egocentric distance were examined. In the triangulation-by-pointing procedure, the observer views a target at some distance and, with eyes closed, attempts to point continuously at the target while traversing a path that passes by it. In the triangulation-by-walking procedure, the observer views a target and, with eyes closed, traverses a path that is oblique to the target; on command from the experimenter, the observer turns and walks toward the target. Two experiments using pointing and 3 using walking showed that perceived distance, averaged over observers, was accurate out to 15 m under full-cue conditions. For target distances between 15 and 25 m, the evidence indicates slight perceptual underestimation. Results also show that observers, on average, were accurate in imaginally updating the locations of previously viewed targets.

Visual control of steering without course information.

An analysis and experimental evidence are presented that the visual control of steering in following a straight demarcated path is possible even when information about the direction of one's travel (course) is not available. Two likely sources of optical information available to observers are considered, bearing and splay, that might be used for steering under these conditions. In an experiment involving a computer-driven driving simulator, observers attempted to steer a straight path while subjected to lateral perturbing forces. When only bearing and its time derivative, motion parallax, were available, performance fell off as expected with the optical gain of motion parallax as the preview distance of the viewing aperture was varied. When splay angle and its time derivative, splay rate, were added to the display, performance generally improved and remained relatively constant with changing distance of the viewing aperture, as expected because of the constant optical gain of splay rate. Making course information available by adding point features to both displays improved steering performance only in the motion-parallax conditions.

Haptic identification of objects and their depictions.

Haptic identification of real objects is superior to that of raised two-dimensional (2-D) depictions. Three explanations of real-object superiority were investigated: contribution of material information, contribution of 3-D shape and size, and greater potential for integration across the fingers. In Experiment 1, subjects, while wearing gloves that gently attenuated material information, haptically identified real objects that provided reduced cues to compliance, mass, and part motion. The gloves permitted exploration with free hand movement, a single outstretched finger, or five outstretched fingers. Performance decreased over these three conditions but was superior to identification of pictures of the same objects in all cases, indicating the contribution of 3-D structure and integration across the fingers. Picture performance was also better with five fingers than with one. In Experiment 2, the subjects wore open-fingered gloves, which provided them with material information. Consequently, the effect of type of exploration was substantially reduced but not eliminated. Material compensates somewhat for limited access to object structure but is not the primary basis for haptic object identification.

Counterexample to the hypothesis of functional similarity between tactile and visual pattern perception.

In earlier work, the author has demonstrated that tactile pattern perception and visual pattern perception exhibit many parallels when the effective spatial resolution of vision is reduced to that of touch, thus supporting the hypothesis that the two pattern senses are functionally similar when matched in spatial bandwidth. The present experiments demonstrate a clear counterexample to this hypothesis of functional similarity. Specifically, it was found that the lateral masking effect of a surround on tactile character recognition increases when the surround changes in composition from solid lines to dots, whereas for vision, recognition performance goes in the opposite direction. This finding necessitates some modification of the model of character recognition proposed by the author (Loomis, 1990) as it applies to the sensing of raised tactile patterns. One possible modification would be to incorporate, as the initial stage of pattern transformation, the continuum mechanics model for the skin that was developed by Phillips and Johnson (1981b).

Nonvisual navigation by blind and sighted: assessment of path integration ability.

Blindfolded sighted, adventitiously blind, and congenitally blind subjects performed a set of navigation tasks. The more complex tasks involved spatial inference and included retracing a multisegment route in reverse, returning directly to an origin after being led over linear segments, and pointing to targets after locomotion. As a group, subjects responded systematically to route manipulations in the complex tasks, but performance was poor. Patterns of error and response latency are informative about the internal representation used; in particular, they do not support the hypothesis that only a representation of the origin of locomotion is maintained. The slight performance differences between groups varying in visual experience were neither large nor consistent across tasks. Results provide little indication that spatial competence strongly depends on prior visual experience.

Visual space perception and visually directed action.

The results of two types of experiments are reported. In 1 type, Ss matched depth intervals on the ground plane that appeared equal to frontal intervals at the same distance. The depth intervals had to be made considerably larger than the frontal intervals to appear equal in length, with this physical inequality of equal-appearing intervals increasing with egocentric distance of the intervals (4 m-12 m). In the other type of experiment, Ss viewed targets lying on the ground plane and then, with eyes closed, attempted either to walk directly to their locations or to point continuously toward them while walking along paths that passed off to the side. Performance was quite accurate in both motoric tasks, indicating that the distortion in the mapping from physical to visual space evident in the visual matching task does not manifest itself in the visually open-loop motoric tasks.

Similarity of tactual and visual picture recognition with limited field of view.

Subjects attempted to recognize simple line drawings of common objects using either touch or vision. In the touch condition, subjects explored raised line drawings using the distal pad of the index finger or the distal pads both of the index and of the middle fingers. In the visual condition, a computer-driven display was used to simulate tactual exploration. By moving an electronic pen over a digitizing tablet, the subject could explore a line drawing stored in memory; on the display screen a portion of the drawing appeared to move behind a stationary aperture, in concert with the movement of the pen. This aperture was varied in width, thus simulating the use of one or two fingers. In terms of average recognition accuracy and average response latency, recognition performance was virtually the same in the one-finger touch condition and the simulated one-finger vision condition. Visual recognition performance improved considerably when the visual field size was doubled (simulating two fingers), but tactual performance showed little improvement, suggesting that the effective tactual field of view for this task is approximately equal to one finger pad. This latter result agrees with other reports in the literature indicating that integration of two-dimensional pattern information extending over multiple fingers on the same hand is quite poor. The near equivalence of tactual picture perception and narrow-field vision suggests that the difficulties of tactual picture recognition must be largely due to the narrowness of the effective field of view.

Active localization of virtual sounds.

A simple virtual sound display built around a microcomputer and analog hardware is described. The display implements most of the primary cues for sound localization in the ear-level plane. Judging both from informal observations by users and from objective data obtained in an experiment on homing to virtual and real sounds, it is concluded that simple displays like the one described are effective in creating the impression of external sounds to which observers can locomote with ease; in particular, this means that simulation of the direction-dependent spectral shaping effects of the pinnae is not a necessary requirement for extracranial sound localization.

Acquisition of route and survey knowledge in the absence of vision.

The ability of sighted, blindfolded individuals to navigate while walking was assessed in two types of tasks, one requiring knowledge of a route that previously had been navigated and another requiring more complex spatial inference or computation. A computerized measurement system monitored spatial position. The route tasks included maintenance of a heading, distance and turn reproduction and estimation, and turn production. The inferential task required completion of a multisegment pathway by returning directly to the origin. pathways were replicated at two different scales. Measures for the route-knowledge tasks indicated a substantial ability to navigate in the absence of visual cues. Route reproduction performance was particularly accurate despite intrinsic veering tendencies. A substantial increase in error was observed in the pattern-completion task. Errors in pathway completion increased with pathway complexity and were quite similar in the two scales. Correlational data suggested that performance on different route-knowledge tasks reflected differing underlying representations. The completion task led to a high correlation between absolute turn and distance error but had minimal correlations with the route tasks. The data suggest that a survey representation with some degree of scale independence was constructed for use in the pathway completion task.

A model of character recognition and legibility.

This article presents a model of character recognition and the experiments used to develop and test it. The model applies to foveal viewing of blurred or unblurred characters and to tactile sensing of raised characters using the fingerpad. The primary goal of the model is to account for variations in legibility across character sets; a secondary goal is to account for variations in the cell entries of the confusion matrix for a given character set. The model consists of two distinct processing stages. The first involves transformation of each stimulus into an internal representation; this transformation consists of linear low-pass spatial filtering followed by nonlinear compression of stimulus intensity. The second stage involves both template matching of the transformed test stimulus with each of the stored internal representations of the characters within the set and response selection, which is assumed to conform to the unbiased choice model of Luce (1963). Though purely stimulus driven, the model accounts quite well for differences in the legibility of character sets differing in character type, size of character, and number of characters within the set; it is somewhat less successful in accounting for the details of each confusion matrix.