Kramer is walking on a treadmill situated in the center of the CAVE™, a 3-D virtual reality immersive environment operated by the Beckman Institute's Integrated Systems Laboratory (ISL) that features three wall-sized display screens in front and on both sides, and another on the floor. Kramer recovers from his encounter with the simulated automobile and lets collaborator and Beckman Fellow Mark Neider take his turn in this newest addition to the world of virtual reality psychology experiments. Neider also successfully crosses the street but he, too, wants to experience the sensation of standing in the path of oncoming traffic.
"You've been hit. Better luck next time." The voice, like the cars, street, and building on the screens is artificial - an AT&T software version of human sound. The images of moving cars were downloaded from the Internet and the street and building images are from pictures of University campus locales taken by Jim Crowell, ISL's triple threat experiment designer/psychologist/ computer programmer.
Crowell, Kramer, and Neider are in the CAVE this mid-August afternoon fleshing out the final form of an experiment that is groundbreaking in its design and in its research focus. It also marks the first time the CAVE has ever been used for an experimental study, said ISL Director Hank Kaczmarski.
The meeting in the Beckman basement home of the CAVE is the culmination of a seven- to eight-month long collaborative process between Crowell and Kaczmarski from ISL and Kramer and Neider, with input from Beckman faculty member Jason McCarley. The study being done by Neider and Kramer involves the topic of pedestrian distraction during traffic situations.
Kaczmarski said designing for experiments almost always includes incorporating something novel, so installing a modified treadmill into the CAVE for the pedestrian study "was a pretty typical process."
However, that doesn't make the pedestrian/ treadmill experiment any less unique, Kaczmarski said. "This has never been done before, anywhere."
Kramer has been a pioneering researcher in the area of driver distraction, with several studies done in Beckman's driving simulator, located just down the hall from the CAVE. He said recent news stories about accidents on campus and around town involving pedestrians prompted this new project that will, like the driver studies, focus on the divided attention of test subjects who are using devices like cell phones.
"We thought that maybe we needed to go beyond driver distraction and start examining pedestrians and how distracted they get as they are crossing busy streets and whether or not it had implications for what happened to them," Kramer said.
The initial run-through session had Kramer, Neider, and Crowell brainstorming on final revisions to the experiment. After taking his turn on the treadmill, Kramer asks Crowell if the speed of the cars can be varied.
"We are kind of pushing the rendering power," Crowell cautions, concerned that increasing the complexity of the simulation might affect the ability to display the cars' movements smoothly.
"Are we able to record all the variables with respect to the simulated automobiles, walking speed, whether you're hit or not hit, where you start relative to where the car is and what the speed is?" Kramer asks.
Yes, they know the speed of the cars and can estimate the speed of the pedestrians, Crowell says.
"What about the distance of a car, say, when they enter the roadway?" Neider asks.
"Sure." The experiment's computer program can even save the test subjects' entire time course at half second or one second intervals throughout the whole course, Crowell adds.
Over the span of the next half-hour, the three discuss the experiment, including issues such as whether to use iPods and if so, what kinds of music (none know what kinds of music "the kids" are listening to so Neider suggests having grad students pick the tunes), or whether to have partners for test subjects if phone usage is studied.
Neider ticks off some of the possible distractions for the pedestrian test subjects: "You could have people just listening to music, you could have them listening to a podcast, or there is the possibility of a different kind of listening going on because they could be watching a TV show on an iPod. Or they could be talking on a cell phone."
Kramer takes things a step further. "We need to buy an iPhone," he says, drawing laughs.
The session ends with Kramer and Neider expressing satisfaction with the design; a little more tweaking and the experiment will be ready for student test subjects in September.
Turning a research idea into the reality of an experimental setting in the CAVE required months of back-and-forth discussions, numerous e-mail exchanges, and ideas that were suggested, cast aside or accepted. The run-through in August with Neider and Kramer was a sort of test-drive in order for Crowell to show the experiment's design to the researchers. While Crowell handled the computer and software requirements, it was up to Kaczmarski to integrate the treadmill into the experiment.
Kramer said there have been studies with treadmills that go at a certain speed but this one operates manually to recreate real-world walking conditions. Kaczmarski modified the treadmill by taking off the display on top and removing a magnet underneath that recorded the number of rotations; he then installed eight magnets that send signals to the PC cluster that powers the CAVE.
The CAVE lab, which had been moved from the third floor to its new home in the basement earlier in 2007, has been used as a testbed for experiments that would eventually be done in the Cube. Kaczmarski said the Cube has become so popular that ISL was forced to start using the CAVE as a facility for experiments, the first being this pedestrian distraction study.
Crowell, who has a Ph.D. in Cognitive Psychology from Cal-Berkeley, is just one of ISL's staff members who aid research efforts in human multimodal perception and cognition using the lab's advanced technologies.
Crowell's role for the past five-and-a-half years has been to assist researchers like Kramer and Neider in setting up experiments.
Crowell is interested in the research side of psychology, but for him the truly interesting part of the equation is the experiment itself.
"The part I like best is the setting up of the experiments," Crowell said. "I have always liked programming and the experiment design part.
"I never so much liked writing things up afterward," he added with a laugh.
Crowell said his background in psychology helps him help the researchers in a number of ways.
"The first and the most obvious way is that when they say something I usually understand immediately what they are talking about, while someone who didn't have the background might spend half an hour figuring it out," Crowell said. "Also, if there is a potential problem in a particular design, I'm fairly likely to be able to spot it, depending on how close it is my own experience."
Neider said Crowell is "genuinely interested in the experiments we are running" while Kramer said ISL is lucky to have someone with backgrounds in both programming and psychology.
"That's ideal to have somebody with both sets of skills. That's pretty rare," Kramer said.
"And Hank's great with hardware." Neider said the pedestrian distraction study had its origins almost a year before when he, Kramer, and McCarley were "throwing out ideas" for new research topics involving divided attention. Kramer said the number of recent accidents on campus involving pedestrians and vehicles seemed like a good topic, both for study and for potentially helping with prevention.
With cell phone usage a necessary fact of life for most people, and music players like iPods a common sight on city streets, the topic of pedestrian distraction is one that is just starting to reach the public consciousness. In New York a bill has been proposed to ban listening to iPods or talking on a cell phone while crossing the street.
"That's where I'm from and it's a wild world there," Neider said. "But there are times when the driver is doing his job and somebody just walks out in front of a car and it happens. There is a big debate starting to grow about this right now, so we are hoping to provide some informed research."
Neider has done mostly theoretical work in the area of visual cognition, specifically in terms of visual search, but this project involves more applied work than his past research. He said the study has multiple goals, and one of them is to extend basic theory regarding attention. Neider said that much of what is known about cognition and attention comes from inferences made from experiments with simple, artificial tasks.
"The problem with that is that the real world is much different than sitting in front of a computer looking at things like a T and an L, which is a classic visual search paradigm," he said. "So the idea here was to start looking at attention, and in this case divided attention because this is a divided attention task, and how what we know about attention from simple tasks translates into a much more realistic scenario where there are a lot of things competing for attention and the task is much more complex."
Thanks to the capabilities of the CAVE and the experiment's designers, the pedestrian distraction study will offer a setting that is more naturalistic and have much more realistic tasks for test subjects to perform. Kramer said the pedestrian study will seek to learn detailed information about how pedestrians negotiate traffic while performing other tasks.
"We want, first of all, to understand the extent to which people can walk safely and navigate busy streets and do other things like listen to an iPod or talk on a cell phone and so forth," Kramer said. "And we want to quantify what the costs are in terms of walking speed or how closely they get to vehicles, when they step off the curb, and how they judge gaps in traffic because we do that all the time."
One of the places on campus that pedestrians and drivers of vehicles are asked by flashing lights and signs to regard each other with an eye toward safety is the crossing on Springfield between Wright and Mathews. That crossing will look familiar to test subjects in the CAVE who are participating in the pedestrian distraction study because Crowell said that is the crossing he had in mind when he was putting together the experiment. The building that is visible to test subjects in the CAVE isn't the nearby Grainger Engineering Library to the south, however, or the Digital Computer Laboratory to the north. Too many trees and other obstacles surrounding them for a clean image, Crowell said.
When he was putting together the experiment Crowell took a walk around campus to scout useful scenes to photograph for the images required on the display screens. He found his building in the new (and as yet treeless or shrub-less) addition to the Micro and Nanotechnology Laboratory just south of Beckman. The virtual crossing has a black asphalt look with wide white stripes just like the one on Springfield, but doesn't include that crossing's flashing lights. No help for the distracted pedestrians, Crowell says.
While the CAVE is now open for business for other perceptual psychology experiments, the first such study done in it won't likely be the last for this line of research. Kramer and Neider discussed other potential experiments involving the CAVE and the treadmill, such as studies involving older adults.
"I would feel pretty confident in saying we will get more than one paper from this," Neider said.
