Visual discrimination thresholds for time to arrival

A Case for Raising the Camera: A Driving Simulator Test of Camera-Monitor Systems

OBJECTIVE

This experiment provides a first-of-its-kind driving-simulator study to investigate the feasibility of camera-monitor systems (CMS) with displaced side-mounted cameras in sedans.

BACKGROUND

Among the increasing number of studies investigating the replacement of side-mounted rearview mirrors with CMS, the placement of side-mounted cameras has been largely neglected. Moreover, user preferences with respect to camera placement have not been validated in a driving simulator. Past research merely has shown that the vertical camera position can affect distance perception.

METHOD

In a driving simulator experiment, we investigated the effects of rearward camera placement on driver acceptance and performance. Thirty-six participants performed multiple lane changes in a last safe-gap paradigm. The camera position, ego-velocity, and velocity of the approaching vehicle varied across the experiment.

RESULTS

The results suggest a clear preference for a high rearward perspective, whereas participants disliked the lower viewpoint. However, these stark differences were only marginally mirrored in lane change performance. Average safety margins tended to decrease and their variation tended to increase for the low camera position.

CONCLUSION

Even if the impact of the camera position on driving behavior seems to be small in sedans, driver expectations show clear-cut preferences. When designing CMS, this should be taken into account, as these preferences could promote the use of CMS and thus their positive impact on safety.

APPLICATION

Designers should place side-mounted cameras as high as possible to increase acceptance of CMS. Low camera positions are not recommended, as they might decrease safety margins and are not appreciated by drivers.

Wider is better but sharper is not: optimizing the image of camera-monitor systems

The replacement of rear-view mirrors with camera-monitor systems introduces new opportunities for design, such as altering the image quality and the rearward field-of-view. We investigated how the image quality and field-of-view might affect the distance and time-to-contact estimation of other vehicles. Eighty-six subjects estimated either their egocentric distance to a stationary vehicle (Experiment I) or the time-to-contact to an approaching vehicle (Experiment II). Throughout the experiments, the pixel density and either the field-of-view or the viewing condition varied. A larger field-of-view increased distance estimation accuracy and confidence. Reduced pixel density led to larger estimates. In contrast, reduced pixel density and simulated dirt shortened time-to-contact estimates. This is compatible with a safety strategy applied under conditions of impaired vision. Moreover, a limited benefit was observed for higher pixel densities. Therefore, camera-monitor systems with large field-of-view and a pixel density of around 300 ppi could ensure accurate TTC and distance estimation. Practitioner summary: A camera's field-of-view and image quality are important parameters for camera-monitor systems. In two experiments, we investigated the effects of these two parameters on rearward distance and time-to-contact estimation. Whereas a larger field-of-view improved distance estimation accuracy, increasing the pixel density had a limited effect in the estimation of time-to-contact.

Estimating time-to-contact when vision is impaired

Often, we have to rely on limited information when judging time-to-contact (TTC), as for example, when driving in foul weather, or in situations where we would need reading glasses but do not have them handy. However, most existing studies on the ability to judge TTC have worked with optimal visual stimuli. In a prediction motion task, we explored to what extent TTC estimation is affected by visual stimulus degradation. A simple computer-simulated object approached the observer at constant speed either with clear or impaired vision. It was occluded after 1 or 1.5 s. The observers extrapolated the object’s motion and pressed a button when they thought the object would have collided with them. We found that dioptric blur and simulated snowfall shortened TTC-estimates. Contrast reduction produced by a virtual semi-transparent mask lengthened TTC estimates, which could be the result of distance overestimation or speed underestimation induced by the lower contrast or the increased luminance of the mask. We additionally explored the potential influence of arousal and valence, although they played a minor role for basic TTC estimation. Our findings suggest that vision impairments have adverse effects on TTC estimation, depending on the specific type of degradation and the changes of the visual environmental cues which they cause.

The Ups and Downs of Camera-Monitor Systems: The Effect of Camera Position on Rearward Distance Perception

OBJECTIVE

This study investigates the effects of different positions of side-mounted rear-view cameras on distance estimation of drivers.

BACKGROUND

Camera-monitor systems bring advantages as compared to conventional rear-view mirrors, such as improved aerodynamics and enlarged field-of-view. Applied research has mainly focused on the comparison between cameras and mirrors or on positioning of in-vehicle monitors. However, the positioning of the exterior camera awaits investigation given that the perspective of the observer at does affect depth perception at large.

METHOD

In two experiments, a total of 50 students estimated metric distances to static vehicles presented in realistic or 3D-rendered pictures. The pictures depicted the rearward scene of a car following the driver as viewed through a camera at varying vertical and horizontal positions. The following vehicle's size and environmental information varied among conditions and experiments.

RESULTS

Lower camera positions led to distance overestimation and higher positions to underestimation. The effect increased as the distance to the following vehicle decreased. Moreover, larger vehicles led to stronger distance underestimation, especially in low camera positions. Interestingly, the main effect of camera position disappeared when the ego-vehicles' back was visible.

CONCLUSION

Different rearward viewpoints affect distance estimation of drivers, especially in close distances. However, a visible reference of one's own vehicle seems to mostly compensate this effect.

APPLICATION

In general, the rear-view camera should be mounted rather higher and to the front of the vehicle. Also, the vehicle's back should always be visible. Low camera positions are not recommended.

Temporal Audiovisual Motion Prediction in 2D- vs. 3D-Environments

Predicting motion is essential for many everyday life activities, e.g., in road traffic. Previous studies on motion prediction failed to find consistent results, which might be due to the use of very different stimulus material and behavioural tasks. Here, we directly tested the influence of task (detection, extrapolation) and stimulus features (visual vs. audiovisual and three-dimensional vs. non-three-dimensional) on temporal motion prediction in two psychophysical experiments. In both experiments a ball followed a trajectory toward the observer and temporarily disappeared behind an occluder. In audiovisual conditions a moving white noise (congruent or non-congruent to visual motion direction) was presented concurrently. In experiment 1 the ball reappeared on a predictable or a non-predictable trajectory and participants detected when the ball reappeared. In experiment 2 the ball did not reappear after occlusion and participants judged when the ball would reach a specified position at two possible distances from the occluder (extrapolation task). Both experiments were conducted in three-dimensional space (using stereoscopic screen and polarised glasses) and also without stereoscopic presentation. Participants benefitted from visually predictable trajectories and concurrent sounds during detection. Additionally, visual facilitation was more pronounced for non-3D stimulation during detection task. In contrast, for a more complex extrapolation task group mean results indicated that auditory information impaired motion prediction. However, a post hoc cross-validation procedure (split-half) revealed that participants varied in their ability to use sounds during motion extrapolation. Most participants selectively profited from either near or far extrapolation distances but were impaired for the other one. We propose that interindividual differences in extrapolation efficiency might be the mechanism governing this effect. Together, our results indicate that both a realistic experimental environment and subject-specific differences modulate the ability of audiovisual motion prediction and need to be considered in future research.