Our bicycling simulator research has two main objectives. One is to better understand the factors that put children at risk for getting hit by motor vehicles when crossing intersections, and the other is to advance the control techniques for creating dynamic events and interactive agents in virtual environments. These two objectives are tightly intertwined in our multidisciplinary research program. Designing experiments to study children’s road-crossing behavior leads to advances in computational techniques, and advances in computational techniques open up new possibilities for studying children’s road crossing behavior.
The Bicycling Simulator
The bicycling simulator provides a unique resource for investigating how children and adults negotiate traffic-filled roadways.
The primary instrument for our work is an immersive, interactive bicycling simulator. The simulator consists of an actual bike mounted on a stationary frame that sits in the middle of three 10 ft wide x 8 ft high screens placed at right angles relative to one another, forming a three-walled room. Three Projection Design F1+ projectors are used to rear-project an image of size 1280 x 1024 pixels onto each of the screens, providing participants with 270 degrees of non-stereoscopic immersive visual imagery. The viewpoint of the scene is adjusted for each rider’s eye height.
The bicycle is instrumented to record the steering angle of the front wheel and the speed of the rear wheel. Steering angle and wheel speed measures are combined with virtual terrain information to render the graphics corresponding to the rider’s real-time trajectory through the virtual environment. The rear wheel is also mated to a friction-drive flywheel.The flywheel is connected to a torque motor, which generates an appropriate dynamic force taking into account rider and bicycle mass and inertia, virtual terrain slope, ground friction, and wind resistance.This provides for a riding experience that realistically incorporates many real-world dynamics of bicycling.
Computationally, the system is a distributed environment hosted on seven PCs connected via a network. The simulation software is divided into motion control for dynamic objects and animation/graphics rendering. The simulation engine (Hank) computes position and orientation for each dynamic object (vehicle, virtual or human rider) on each step of the simulation. This information is then transmitted to graphics PCs, where it is processed by the Visualizer application to update the scene graphs and render a corresponding image for each of the screens.
Dual-Networked Interactive Simulation
Our interactive simulation technology will allow two real riders (i.e., a child and friend or a child and parent) to cross roads together via the medium of two connected bicycling simulators. Each rider will control their motion and be surrounded by graphically rendered images that show a first-person view of their motion through the environment. Avatars will represent rider motions to provide a realistic experience of riding with another person.
Children and adolescents engage in many daily physical activities in the company of others. They frequently cross streets, ride bikes, and drive cars with peers and parents. Numerous studies indicate that peers exert a strong influence on risk taking, particularly in late childhood and early adolescence. Because of the inherent danger in examining social influences on risk taking in situ, we know very little about how social influences operate while children and adolescents are actually engaged in potentially risky physical activities. The goal of our multi-disciplinary research project is to develop interactive simulation technology as a safe and effective way to study social influences on child cyclists’ road-crossing behavior. Our past work has shed light on perceptual-motor risk factors for car-bicycle collisions by examining how children cross intersections using an interactive, immersive bicycling simulator. Our future work will significantly advance understanding of social risk factors for car-bicycle collisions by developing technology to study children’s interactions with virtual avatars and agents while riding with them through our virtual environment.
Our computational work focuses on developing simulation software to create complex yet robust scenarios for our bicycling experiments and on developing sophisticated techniques for visualizing and analyzing the data from our experiments. To read more about the computational foundations of this work, click here.
Below we describe our inidvidual projects on how children and adults cross roads in our virtual environment. Click on the titles of projects to download pdfs of the published research.
This investigation examined how children and adults negotiate a challenging perceptual-motor problem with significant real-world implications— bicycling across two lanes of opposing traffic. Twelve- and 14-year-olds and adults rode a bicycling simulator through an immersive virtual environment.
Participants crossed intersections with continuous cross traffic coming from opposing directions. Opportunities for crossing were divided into aligned (far gap opens with or before near gap) and rolling (far gap opens after near gap) gap pairs. Children and adults preferred rolling to aligned gap pairs, though this preference was stronger for adults than for children. Crossing aligned versus rolling gap pairs produced substantial differences in direction of travel, speed of crossing, and timing of entry into the near and far lanes. For both aligned and rolling gap pairs, children demonstrated less skill than adults in coordinating self and object movement. These findings have implications for understanding perception–action– cognition links and for understanding risk factors underlying car– bicycle collisions
|Child rider with the virtual peer.||
How do peers influence children's road-crossing behavior? Because children often cross roads with other children, peers likely constitute an important influence on children's road-crossing behavior. We examined whether 10- and 12-year-olds' gap choices are influenced by riding with a peer who exhibits risky vs. safe road-crossing behavior.
To study this problem, we created a virtual peer bicyclist to ride with the child through our virtual environment. Following the warm-up session, children met a same-sex virtual peer, who informed them that after they arrived together at each intersection, children would first watch the peer cross the intersection and then they would cross the intersection on their own. Children rode with the peer for the first six intersections, and then crossed the last six intersections alone. The peer always took a 3.5 s gap in the risky peer condition and a 5.5 s gap in the safe peer condition. We found that children who rode with a risky peer were more likely to cross 3.5 and 4.5 s gaps than were children who rode with a safe peer. These gaps are ambiguous in the sense that they are neither too small to cross nor are they easily crossable. Thus, it appears that watching a peer take tighter gaps may lead children to take tighter gaps themselves.
|Image of child rider in the bicycling simulator.||
Are there age differences in how 10- and 12-year-old children and adults bicycle across traffic-filled intersections? We addressed this question by examining whether gap sizes and movement timing differed between children and adults.
Children and adults rode our bicycling simulator through a virtual environment consisting of a straight, residential street with six intersections. Participants faced cross traffic from their left-hand side and waited for gaps they judged were adequate for crossing. The results clearly showed that relative to adults, children's gap choices and road-crossing behavior were less finely tuned. Children and adults chose the same size gaps and yet children ended up with less time to spare when they cleared the path of the approaching car. In fact, the margin for error at this critical juncture was very small, particularly for 10-year-olds. Why did children end up with less time to spare than adults? When we looked back at what the children and adults were doing when they initiated their movement to cross, we see that children delayed initiation of crossing relative to adults. This resulted in less time to spare when children clear the path of the approaching car. These differences in how children and adults initiate movement suggest that immature perceptual-motor skills may play a role in putting children at greater risk for car-bicycle collisions.
Video of adult rider in interception task.
Our recent work has examined why children delay initiation of movement when crossing roads relative to adults. One possibility is that children have less precise control over the synchronization of self and object movement. Hence, they may delay in initiating movement to allow a greater safety margin between themselves and the lead vehicle in the gap. To study developmental changes in the synchronization of self and object movement, we created a gap interception task.
Ten- and 12-year-old children and adults rode an actual bicycle through our virtual environment. At each of 12 intersections, participants attempted to pass between two moving red blocks without stopping(see video). Block motions were timed such that participants would arrive early or late in the target gap if they maintained constant speed. This meant it was sometimes necessary for participants to speed up or slow down in order to intercept the gap. Our primary goal was to determine whether children's time to spare (relative to the rear car of the gap) was less than that for adults. This would suggest that time-to-spare deficiencies observed in previous studies were not solely due to issues with poor movement preparation strategies or difficulties in go no-go decision-making.
As in our road-crossing work, children had less time to spare than adults when they intercepted the blocks. Children also exhibited significantly more variability in the amount of time they had to spare. Thus, even though children did not have to select the gap to cross or initiate movement from a stop, they still timed their movements less skillfully than did adults. We also found that children’s approach profiles were more erratic than those of the adults, with more pronounced corrections in speed as they approached the intersection. In sum, although the patterns of interceptive actions were similar in children and adults, children’s interceptive actions were less finely tuned than those of adults.
|Child rider waiting to cross intersection.||
How does experience with performing a task lead to change in decisions and actions? We addressed this question by examining how experience with our road-crossing task helps children bring their gap choices and crossing behavior more tightly in line with perceptual information.
Ten- and 12-year-old children and adults bicycled across 12 intersections with continuous cross-traffic coming from their left-hand side. We manipulated traffic density to examine how the experience of operating near the limit of the perceptual-motor system affected later gap choices and movement timing. In the control condition, children and adults encountered randomly ordered gaps ranging from 1.5 to 5 s at all intersections. In the high-density condition, children and adults encountered a set of intersections with high-density traffic sandwiched between sets of intersections with randomly ordered gaps ranging from 1.5 to 5 s. Thus, the first four and last four intersections were the same for both groups, but the middle four intersections differed. This allowed us to directly examine the extent to which change in gap choices over the session was due to general experience with crossing intersections or to specific experience with high-density traffic.
The results clearly revealed changes in gap choices and movement timing over the course of the session. Children and adults chose smaller gaps to cross at the last set of intersections than at the first set of intersections. The tendency to take smaller gaps during the last set of intersections was also influenced by the type of previous experience. Participants who experienced high-density traffic prior to the last set of intersections also accepted more tight 3 s gaps during the last than the first set of intersections. We also found changes in movement timing over the course of the session, particularly for the 10-year-olds. The 10-year-olds had significantly more time to spare between themselves and the oncoming car during the last four intersections than during the first four intersections. By the last four intersections, they had increased their time to spare by an average of 25% (.44 s). More recent work shows that 10-year-olds exhibit the same kinds of improvement in movement timing over the course of the session in our gap interception task. Together, these results clearly show that 10-year-olds are transitional with respect to the ability to coordinate self and object movement in our road-crossing task.