BerthozBBCFFHKMMNPRSSTVvVW2013 3 A Berthoz W Bles HH Bülthoff BJ Correia Gracio P Feenstra N Filliard R Huhne A Kemeny M Mayrhofer M Mulder HG Nusseck P Pretto G Reymond R Schlüsselberger J Schwandtner H Teufel B Vailleau MM van Paassen M Vidal M Wentink 2013-05-00 3 43 265 276 IEEE Transactions on Human-Machine Systems Advanced driving simulators aim at rendering the motion of a vehicle with maximum fidelity, which requires increased mechanical travel, size, and cost of the system. Motion cueing algorithms reduce the motion envelope by taking advantage of limitations in human motion perception, and the most commonly employed method is just to scale down the physical motion. However, little is known on the effects of motion scaling on motion perception and on actual driving performance. This paper presents the results of a European collaborative project, which explored different motion scale factors in a slalom driving task. Three state-of-the-art simulator systems were used, which were capable of generating displacements of several meters. The results of four comparable driving experiments, which were obtained with a total of 65 participants, indicate a preference for motion scale factors below 1, within a wide range of acceptable values (0.4–0.75). Very reduced or absent motion cues significantly degrade driving performance. Applications of this research are discussed for the design of motion systems and cueing algorithms for driving simulation. no notspecified http://www.kyb.tuebingen.mpg.de/ published 11 15017 15422 PrettoBRB2012 3 P Pretto J-P Bresciani G Rainer HH Bülthoff 2012-10-00 1 1 12 eLife Visual speed is believed to be underestimated at low contrast, which has been proposed as an explanation of excessive driving speed in fog. Combining psychophysics measurements and driving simulation, we confirm that speed is underestimated when contrast is reduced uniformly for all objects of the visual scene independently of their distance from the viewer. However, we show that when contrast is reduced more for distant objects, as is the case in real fog, visual speed is actually overestimated, prompting drivers to decelerate. Using an artificial anti-fog—that is, fog characterized by better visibility for distant than for close objects, we demonstrate for the first time that perceived speed depends on the spatial distribution of contrast over the visual scene rather than the global level of contrast per se. Our results cast new light on how reduced visibility conditions affect perceived speed, providing important insight into the human visual system. no notspecified http://www.kyb.tuebingen.mpg.de/ published 11 15017 15422 5694 3 P Pretto M Ogier HH Bülthoff J-P Bresciani 2009-04-00 2 33 139 146 Computers and Graphics Efficient navigation requires a good representation of body position/orientation in the environment and an accurate updating of this representation when the body-environment relationship changes. Such updating is based on the ability to correctly estimate the speed and amplitude of body displacements. Because navigation in virtual worlds often relies on the sole visual information, we investigated to which extent the size of the field of view (FoV) affects two basic aspects of motion perception: (i) the perceived amplitude of rotations about the body vertical axis (Experiment 1) and (ii) the perceived speed of forward translations (Experiment 2). Concerning the perception of rotation amplitude, we found that visual flow information gives rise to inaccurate and very variable estimations, with a systematic underestimation of rotations larger than 30 degrees. We also found that the accuracy of the estimations does not depend on the size of the FoV and that horizontal FoVs larger than 30 degrees do not improve the performance. Concerning speed perception, central FoVs smaller than 60 degrees gave rise to an underestimation of the visual speed. On the other hand, occluding the central area leaving only peripheral visual information available induced a systematic overestimation of visual speed, even when only the central 10 degrees of vision were occluded. Taken together, these results suggest that large FoVs are not required to estimate the amplitude of visual rotations about the vertical axis of the body, whereas central FoVs of at least 60 degrees are advisable when speed perception relies on visual flow information. no notspecified http://www.kyb.tuebingen.mpg.de/ published 7 15017 15422 3044 3 P Pretto G Monacelli L Gamberini 2003-00-00 2 1 164 175 PsychNology Journal no notspecified http://www.kyb.tuebingen.mpg.de/ published 11 NestiMBRBP2012 7 A Nesti C Masone M Barnett-Cowan P Robuffo Giordano HH Bülthoff P Pretto Paris, France2012-09-00 1 6 Driving Simulation Conference Europe (DSC 2012) Due to limited operational space, in dynamic driving simulators it is common practice to implement motion cueing algorithms that tilt the simulator cabin to reproduce sustained accelerations. In order to avoid conflicting inertial cues, the tilt rate is kept below drivers’ perceptual thresholds, which are typically derived from the results of classical vestibular research where additional sensory cues to self-motion are removed. Here we conduct two experiments in order to assess whether higher tilt limits can be employed to expand the user’s perceptual workspace of dynamic driving simulators. In the first experiment we measure detection thresholds for roll in conditions that closely resemble typical driving. In the second experiment we measure drivers’ perceived realism in slalom driving for sub-, near- and supra-threshold roll rates. Results show that detection threshold for roll in an active driving task is remarkably higher than the limits currently used in motion cueing algorithms to drive simulators. Supra-threshold roll rates in the slalom task are also rated as more realistic. Overall, our findings suggest that higher tilt limits can be successfully implemented in motion cueing algorithms to better optimize simulator operational space. no notspecified http://www.kyb.tuebingen.mpg.de/fileadmin/user_upload/files/publications/2012/DSC-2012-Nest.pdf published 5 15017 15422 6632 7 V Grabe P Pretto P Robuffo Giordano HH Bülthoff Paris, France2010-09-00 81 88 Driving Simulation Conference Europe (DSC 2010) Different solutions are used on driving simulators to provide visual feedback. In this study, we investigated the influence of projection technology and field of view on drivers performance in a slalom driving task. We tested a head mounted display against a curved projection system on our CyberMotion simulator, based on an anthropomorphic robot arm. The results showed that drivers performed significantly better using the projection screen than the HMD. The FoV and the motion simulation did not have a measurable influence on the performance. no notspecified http://www.kyb.tuebingen.mpg.de//fileadmin/user_upload/files/publications/DSC2010_Grabe-et-al_6632[0].pdf published 7 15017 15422 5628 7 P Pretto H-G Nusseck H Teufel HH Bülthoff Monte Carlo, Monaco2009-02-00 121 131 Driving Simulation Conference Europe (DSC 2009) no notspecified http://www.kyb.tuebingen.mpg.de/ published 10 15017 15422 5088 7 P Pretto M Vidal A Chatziastros Monaco2008-02-00 223 235 10th Driving Simulation Conference In the first experiment we investigated the effect of reduced visibility on the produced speed in a driving simulation. Participants were required to drive at a target speed of 90 km/h in different visibility conditions. We found that when realistic fog was simulated, the driving speed was reduced accordingly to the fog density. When a uniform reduction of the image contrast was implemented, no effects were observed on the produced speed. We speculated that fog reduces selectively the visibility of the distant region of the scene and leaves visible only the proximal area that contains high angular velocities. We hypothesized that the perceived speed is then biased by the available raw velocity signals from the visual field. In the second experiment we addressed the question whether the observed behavioral effect has indeed a perceptual origin. In a psychophysical task we asked the participants to estimate the speed of moving scenes when the sight was limited either to the periphery (high angular velocities) o r to the center (low angular velocities) of the field of view. According to our hypothesis, we found that when the central region was occluded, the speed at the periphery was perceived as being higher, and conversely, when the peripheral region was missing the speed at the center was perceived as being lower. We conclude that the speed reduction while driving in fog is due to a non-optimal perceptual compensation for the hidden central region with low angular velocities, which causes an overestimation of the driving speed. no notspecified http://www.kyb.tuebingen.mpg.de/ published 12 15017 15422 4076 7 P Pretto A Chatziastros Paris, France2006-10-00 263 272 Driving Simulation Conference Europe (DSC Europe 2006) no notspecified http://www.kyb.tuebingen.mpg.de/fileadmin/user_upload/files/publications/DSC-2000-Pretto.pdf published 9 15017 15422 4077 7 P Pretto A Chatziastros Dublin, Ireland2006-07-00 1 8 11th International Conference on Vision in Vehicles (VIV 2006) no notspecified http://www.kyb.tuebingen.mpg.de/ published 7 15017 15422 3048 46 L Gamberini P Pretto A Grassi LD Stasi 2004-00-00 2004-00-00 Virtual Environments in Industrial Prototyping no notspecified NestiBPB2012 7 A Nesti K Beykirch P Pretto HH Bülthoff Schramberg, Germany2012-11-00 27 13th Conference of the Junior Neuroscientists of Tübingen (NeNA 2012) Sensory information processes leading to human self-motion perception have been modelled in the past in terms of visual and inertial stimulations and their interactions. The models, validated through many psychophysical experiments, rely on the assumption that our sensitivity to supra-threshold self-motion is not affected by motion intensity. In other words, the relationship between motion stimulus intensity and human sensitivity to motion is assumed to be linear. However, recent studies have shown that this relationship is non-linear, in particular at higher motion intensity. Therefore, the implementation of nonlinearities in the computational models of human motion perception would increase their accuracy over a wider range of motion stimulus intensity. Here we test human sensitivity for sinusoidal yaw rotation in darkness at frequencies of 0.5 Hz and 1 Hz and velocity amplitudes ranging between 0 and 90 deg/s. In a two interval force choice experimental paradigm, subjects undergo two consecutive rotations in the same direction for each trial. One of these movements is repeated unchanged in every trial, while the other systematically varies in amplitude. Subjects are asked to report after each trial which one of the two movements was stronger. An adaptive staircase adjusts the motion for every trial to identify the smallest detectable change in stimulus intensity (differential threshold). Results show a power law relationship between differential thresholds and stimulus intensity, meaning that sensitivity decreases as motion becomes stronger. No frequency effect is observed. These findings are of particular interest for the field of vehicle motion simulation, where knowledge about self-motion perception is widely exploited to overcome the physical limitations of motion-based simulators. Furthermore, the identification of perceptual nonlinearities in multisensory stimulation will guide future work into understanding the neural mechanisms responsible for self-motion perception. no notspecified http://www.kyb.tuebingen.mpg.de/ published -27 15017 15422 NestiBBP2012 7 A Nesti M Barnett-Cowan HH Bülthoff P Pretto Oxford, UK2012-06-21 167 13th International Multisensory Research Forum (IMRF 2012) The restricted operational space of dynamic driving simulators requires the implementation of motion cueing algorithms that tilt the simulator cabin to reproduce sustained accelerations. In order to avoid conflicting inertial cues, the tilt rate is limited below drivers’ perceptual thresholds, which are typically derived from the results of classical vestibular research, where additional sensory cues to self-motion are removed. These limits might be too conservative for an ecological driving simulation, which provides a variety of complex visual and vestibular cues as well as demands of attention which vary with task difficulty. We measured roll rate detection threshold in active driving simulation, where visual and vestibular stimuli are provided as well as increased cognitive load from the driving task. Here thresholds during active driving are compared with tilt rate detection thresholds found in the literature (passive thresholds) to assess the effect of the driving task. In a second experiment, these thresholds (active versus passive) are related to driving preferences in a slalom driving course in order to determine which roll rate values are most appropriate for driving simulators so as to present the most realistic driving experience. The results show that detection threshold for roll in an active driving task is significantly higher than the limits currently used in motion cueing algorithms, suggesting that higher tilt limits can be successfully implemented to better optimize simulator operational space. Supra-threshold roll rates in the slalom task are also rated as more realistic. Overall, our findings indicate that increasing task complexity in driving simulation can decrease motion sensitivity allowing for further expansion of the virtual workspace environment. no notspecified http://www.kyb.tuebingen.mpg.de/ published -167 15017 15422 PrettoBB2010 7 P Pretto J-P Bresciani HH Bülthoff Lausanne, Switzerland2010-08-00 168 33rd European Conference on Visual Perception Perceived speed of moving patterns is lower when visual contrast is reduced. In recent driving simulation experiments, this has been presented as an explanation why drivers might drive too fast in fog. In this simulation, contrast was equally reduced for all objects in the scene, independently of their distance from observer (distance-independent contrast reduction). In fog, however, contrast is reduced more for more distant objects (distance-dependent contrast reduction). We compared the effects of these two types of contrast reduction on both perceived and actual driving speed. In the first experiment twelve participants were presented with pairs of driving scenes—one with clear visibility, one with reduced contrast—and instructed to estimate which scene moved faster. Speed was underestimated with distance-independent contrast reduction but overestimated with distance-dependent contrast reduction. In the second experiment, ten drivers drove at target speeds under clear and reduced visibility conditions without tachometer. Participants increased speed with distance-independent contrast reduction but decreased speed with distance-dependent contrast reduction. These results show how visibility loss can lead to opposite perceptual and behavioural effects, depending on the nature of the underlying visual contrast reduction. In fog, indeed, the visual system fools drivers the safe way, advising them to decelerate. no notspecified http://www.kyb.tuebingen.mpg.de/ published -168 15017 15422 BergerPBB2009 7 D Berger P Pretto HH Bülthoff J-P Bresciani Regensburg, Germany2009-08-00 134 32nd European Conference on Visual Perception Size cues are known to affect the perceived distance between objects. If the physical distance between two objects remains constant, the larger the retinal image of the objects is, the closer to one another they are perceived to be. We tested whether size cues affect the perceived visual speed of self-motion. Ten subjects sitting in front of a panoramic screen (230 × 125° of field of view) were presented with constant-speed translations of the visual scene, which simulated a forward translation of the body. The physical size of the objects present in the scene was systematically varied, and the subjects had to estimate the speed of the translations. We used a 2IFC task, ie, two stimuli were presented successively and the subject had to indicate which one was faster. The experiment had four conditions, with two main factors: object size (big/small) and object familiarity (absolute size known/unknown). Eye and horizon heights were constant throughout all conditions. We observed a main effect of object size, with speed being underestimated for larger objects. Interestingly, this effect was significantly stronger for objects whose absolute size is known. These results suggest a top - down influence of size cues on the perceived visual speed of self-motion. no notspecified http://www.kyb.tuebingen.mpg.de/ published -134 15017 15422 5309 7 P Pretto J-P Bresciani M Vidal HH Bülthoff Utrecht, Netherlands2008-08-00 156 31st European Conference on Visual Perception During translations along the antero-posterior axis, the angular velocity of the visual flow on the retina varies with the retinal eccentricity of the stimulus. We tested how the perceived translation speed is affected by the portion of the retina that is stimulated. Twelve seated subjects were presented with visual translations at constant speed through a volume of random dots. The perceived speed was compared between different field-of-view (FOV) conditions, masks of different sizes being used to occlude either central or peripheral areas of the FOV (230*125 degrees when non-occluded). With central FOVs smaller than 40 degrees (ie, peripheral vision occluded), the visual speed was systematically underestimated, the bias being inversely proportional to the size of the FOV. In contrast, when the central region was occluded and visual flow only presented peripherally, the speed was systematically overestimated. This overestimation was observed even when only 10 degrees of central FOV were occluded. Our results suggest that correct perception of visual speed requires at least 40 degrees of central FOV. no notspecified http://www.kyb.tuebingen.mpg.de/ published -156 15017 15422 5308 7 M Vidal P Pretto Utrecht, Netherlands2008-08-00 70 31st European Conference on Visual Perception While walking, the visual scene analysis provides both structural information about the environment, and self-motion characteristics. These two categories are in fact strongly interrelated. During constant speed translations on a plane, the angular retinal speed varies according to where we look at, and still self-linear speed is perceived as constant which allows the world to appear stable. We studied factors involved in the retinal-to-linear compensation mechanism underlying this perceptual stability. In all experiments we used a speed discrimination task between two gaze directions (12°, 20°, or 28° below horizon). When keeping the eyes static, the compensation was almost perfect if viewed with a full field 86%, whereas it was poorer, but far from null, if viewed with a 6° vertical aperture 74%. When tracking a target on the plane, this compensation improved for both full and limited field. Finally, reducing the visual scene to the tracked target also allowed for good compensation, though worst than with the plane. These findings are discussed with regard to inverted structure-from-motion mechanisms. no notspecified http://www.kyb.tuebingen.mpg.de/ published -70 15017 15422 4891 7 P Pretto M Vidal A Chatziastros Tübingen, Germany2007-07-00 154 10th Tübinger Wahrnehmungskonferenz (TWK 2007) During linear self-motion at constant speed, the retinal speeds of stationary objects vary as a function of their declination angle (the angle between the line of sight and the horizontal plane). Nevertheless, when we move in our environment, we do not feel that different places move at different speeds: a compensation mechanism is thought to mediate between angular velocity and perceived linear speed so that velocity constancy is achieved. In a recent study [1] it has been shown that the perceived speed is altered when driving with a reduced fieldof- view (FOV). The explanation proposed in that study leads us to the hypothesis that, when moving at constant speed, humans might not be able to compensate for the different velocity signals coming from various declination angles when only a limited portion of the visual field is visible. Here we tested this hypothesis using a Virtual Reality (VR) setup that provides a 230×125 (H×V) FOV. We measured the visual perceived speed at eye-height (1.7m) while simulating fast walking speeds on a virtual open field. We manipulated the FOV (full field vs. limited field corresponding to an aperture of 40×6) and the gaze declination angle (12, 20 and 28 degrees), corresponding to positions on the plane located at a distance of 8, 4.7, and 3.2 m, respectively. We used a two alternative forced choice (2AFC) with constant stimuli method in a 2×3 within subjects design. We tested eight different speeds ranging from 0.67 to 6 m/s. The reference stimulus appeared always in the intermediate declination angle at the speed of 2 m/s. A fixation cross appeared at the desired declination angle 500 ms before each stimulus. At every trial, subjects had to select which of the two presented stimuli indicated a faster linear forward speed. The results of four observers show that when looking with a different declination angle in the test, the perceived speed appeared either higher or lower than the reference speed. This effect was accentuated in the limited FOV condition, suggesting that limiting the FOV impairs the compensation mechanism. Interestingly, while two observers could not fully compensate for the perceived retinal speed even within a full FOV condition, the other two showed a reliable over-compensation independently of the FOV. This indicates that a veridical speed estimation cannot be achieved in VR and with limited FOV and that speed estimation is not independent of gaze direction. no notspecified http://www.kyb.tuebingen.mpg.de/ published -154 15017 15422 3853 7 P Pretto A Chatziastros Tübingen, Germany2006-03-00 162 9th Tübingen Perception Conference (TWK 2006) Optic flow affects the walking speed during human pedal locomotion: when the ground texture is moved in a direction opposite to the walking direction, people slow down despite noticeable changes in their physical effort. We hypothesized that in driving conditions, where no direct matching between physical effort and produced speed exists, the role of the optic flow should be even more pronounced. Similarly, fog is supposed to reduce the perceived driving speed, by a reduction of the contrast of the visual scene. This, subsequently, should lead to a speeding during fog [1]. We aimed to replicate these results using a large field-of-view, a real road trajectory and a realistic exponential fog model. We used a driving simulation which was back-projected on a large fronto-parallel projection screen (75 x 70 degrees). Participants were trained to learn the relation between current driving speed (digital speedometer) and amount of optic flow in the image plane. In the experimental phase, we varied environmental condition (clear vs. foggy scene), speed signal (40, 60, and 90 km/h), and road texture motion. By adding motion to the road texture, the apparent driving speed was increased to 150% or reduced to 67% of the actual driving speed with regard to the rest of the environment. After the beginning of the trial, a speed signal indicated the required speed. The participants were instructed to accelerate up to this speed, keep it for five seconds, and terminate the trial by button press. The average speed of the last five seconds of each trial was included into the analysis. We found a significant effect of the motion of the road texture on the produced driving speed. As expected, participants increased the driving speed when the texture motion indicated a slower speed and vice versa. We observed also a significant fog effect which is in opposition to the previous results: when contrast was attenuated by fog the driving speed was reduced. Our study raises serious doubts about the previous interpretation of the relation among contrast reduction, perceived speed and driving behavior. In a realistic driving scenario, fog does not directly lead to speeding. However, the large effect of texture motion, suggests that the optic flow, especially originating from the road ahead, clearly determines the perception of the driving speed, even in foggy conditions. no notspecified http://www.kyb.tuebingen.mpg.de/ published -162 15017 15422 3322 7 P Pretto A Chatziastros Tübingen, Germany2005-02-00 186 8th Tübingen Perception Conference (TWK 2005) Recently, the role of luminance contrast on the perception of velocity has been emphasized by the so called “footstep illusion” [1]: when a grey bar drifts steadily across a background of stationary black and white stripes, it appears to stop and start as its contrast to the background changes. When two bars of unequal grey values drift in parallel, an impression of alternating footsteps of a moving person can be elicited. This change in apparent speed has been attributed to the different instantaneous contrasts of the grey bar against the white and black areas of the background (Contrast Ratio explanation, [1]). We report a series of experiments with further variants of the footstep illusion to test the validity of the Contrast Ratio (CR) explanation. In an adjustment task, we presented a standard footstep illusion together with second comparison stimulus. A total of five participants were required to adjust the contrast of the background of the comparison stimulus in order to match the strength of the standard illusion. The moving bars were presented in six luminance levels in the range between 6 and 103 cd/m2 . For each luminance level we computed the CR by the Weber fraction, following the original study. We compared the CR values of the standard and comparison stimulus and expected to obtain comparable CRs within matched illusion pairs. In the first experiment we found a discrepancy between predicted and obtained CR values. The strength of the illusion was matched at an increased CR, i.e. participants set the background stripes to higher contrast in order to obtain the same illusion. In a second experiment the moving bars were rendered transparent, with a transparency of 50%, and we obtained basically the same results as in the first experiment. In the last experiment we tested whether the footstep illusion can be elicited with isoluminant red/green patterns. Individual isoluminance levels were identified with flickering stimuli. Subjects rated an apparent velocity change of the isoluminant moving bars again by adjusting the contrast of the standard luminance version. The results show clearly that the footstep illusion occurs also with isoluminant patterns. Furthermore, we found that the adjusted CR was always near the maximal possible value, suggesting that participants matched the isoluminant illusion to the strongest available luminance illusion. The present data suggest that foreground and background stimuli contribute differently to the footstep illusion, a finding which is not predicted by the current Contrast Ratio account. Most notably, the existence of the illusion with isoluminant stimuli shows the clearest limitation of the Contrast Ration explanation. That the strength of the isoluminant footstep illusion was comparable to the luminance version, suggests a crucial role of chromatic input in an adequate explanation of this phenomenon. no notspecified http://www.kyb.tuebingen.mpg.de/ published -186 15017 15422 Pretto2012 10 P Pretto Pretto2012_2 10 P Pretto Pretto2011 10 P Pretto Pretto2009_2 10 P Pretto Pretto2009 10 P Pretto Pretto2008 10 P Pretto A Chatziastros 3506 10 P Pretto 3046 10 P Pretto 3047 10 P Pretto