BLIND PEDESTRIANS AND QUIETER VEHICLES: HOW ADDING ARTIFICIAL SOUNDS IMPACTS TRAVEL DECISIONS

摘要

his research examined the influence “quiet” (e.g.,hybrid and electric) vehicles may have on the ability of blindpedestrians to perform common orientation and mobility tasksunder low vehicle speed conditions. The research involvedblind participants detecting forward approaching vehicles andapproaching backing vehicles, deciding whether a vehiclecoming from behind or from the front but across anintersection would continue to go straight or turn across theintended path of travel of a pedestrian seeking to cross a street(i.e., a pathway discrimination task), and taking parallel andperpendicular alignment from passing traffic. Participantsincluded some with normal hearing and some with impairedhearing.Testing was conducted on a public roadway and aparking lot in Kalamazoo, Michigan under ambient soundconditions consistent with a typical urban travel environment.Conditions involved evaluating internal combustion engine(ICE) Chevrolet Malibu and a set of hybrid Chevrolet Voltscapable of operating in a “quiet” mode (referred to as ElectricMode or EM) or operating in EM but augmented with one offive different artificially-generated sounds emanating out of afront-bumper mounted speaker.All of the artificial sounds generally performed betteragainst the baseline the Chevrolet Volt EM than the ChevroletMalibu. This suggests that, to some extent, putting any one ofthese artificial sounds on a hybrid or electric vehicle mayimprove pedestrian performance on the measures examinedrelative to not adding any sound at all. One sound (sound 5)did not outperform against the Chevrolet Malibu in anymeasure and had the fewest instances of outperforming theChevrolet Volt EM. Of the remaining 4 sounds, two soundsoutperformed both the Chevrolet Malibu and the ChevroletVolt EM on several measures. Against the Chevrolet Malibu,sound 2 was slightly better in detection distance and crossingmargin while sound 4 was better in the path discriminationtasks. The two sounds were equivalent on the alignment tasks.The pathway discrimination task reflects one of the morepotentially threatening situations in which a blind pedestrianmight encounter a quiet vehicle (e.g., turning to cross thepedestrian’s path). Sound 4 performed much better than sound2 on this measure, making it the most effective of all theartificial sounds examined. While these two sounds wereequitable in the right-straight task, sound 4 showed almost halfas many missed vehicle surges (i.e., forward movement from astop) and ¼ the rate of missed paths and incorrect judgments.Vehicle sound condition did not impact participants’alignment. Normal hearing participants performedsignificantly better than hearing impaired participants on thistask, but not as well as would be expected based on previousdata [1].These results support the potential for artificiallygeneratedsounds to improve the ability of blind pedestrians todetect approaching vehicles relative to what is being achievedwith ICE vehicles. Regression analysis of the detection datasupports previous results that sound energy in the 500 to 1000Hz range is important for detection. However, the analysisindicates it is not that energy in this region that makes thesignal more noticeable, but that energy in this region in theambient environment hinders detection. Previous findings in low ambient conditions showing a predictive value for theamplitude modulation of an artificial sound were not supportedin these data.