Programs & Labs
Winter 2007
Haptic feedback research explores driver cognition issues
Mike Manser
Only a few decades ago, catastrophic failure of critical components was a common cause of automobile crashes. Broken tie-rods sent cars hurtling into ditches; brake systems failed suddenly under the strain of mountain driving. Today, improvements in automotive engineering have greatly reduced the number of crashes caused by mechanical failure while driving. So why are so many people still losing their lives every year in automobile crashes?
Statistics show that more than 40,000 Americans die every year on the road—that adds up to more than 650,000 deaths since 1990. And if current trends continue, driving-related incidents will become the third leading cause of death in this country by 2020, up from ninth in 1990.
"These numbers are simply staggering," says human factors researcher Michael Manser of the ITS Institute's HumanFIRST Program.
One explanation for this apparent paradox is that while engineers have been very successful in making vehicles safer, drivers remain largely unchanged. Today, driver error may be as significant as mechanical failures. And with the myriad electronic controls and options present in today's high-tech vehicles, not to mention cellular phones and other personal communication devices that demand drivers' attention, the potential impediments to good driving performance may be increasing.
All drivers, no matter how experienced, are subject to natural limits of human behavior, cognition, and perception. But just as technology can help overcome physical limitations, it can also help address the perceptual and cognitive biases that often lead to less than optimal driving performance. The potential solution, says Manser, is using technologies that support—rather than impede—good driving practices.
Manser, who is currently interim director of the HumanFIRST Program, has been involved in a range of research projects focused on issues of driver performance since joining the University of Minnesota in 2002. For the past four years, he has been one of the primary researchers in a collaborative effort with Nissan Motor Company of Japan aimed at evaluating a new driver-assistive system that helps improve driver perception of lead vehicle status changes.
The Nissan system uses a haptic (touch-based) feedback mechanism attached to the accelerator pedal to provide variable resistance depending on how close the driver's vehicle is to a lead vehicle. The closer the vehicle gets to the lead vehicle, the more the pedal pushes back against the driver's foot. Forward-looking range sensors are able to sense changes in distance much more accurately and quickly than the human eye, and relay these changes instantly even if the driver's attention is elsewhere.
Car following tests of the haptic feedback system were carried out on the MnROAD test track.
Although humans are endowed with highly evolved senses of hearing and touch, we rely almost exclusively on vision when we get behind the wheel. In this context, says Manser, haptic feedback systems are interesting because they exploit a relatively underused information channel that may not compete with the many visual cues that drivers already have to process. ITS Institute researchers have experimented with a number of other haptic feedback applications in the past, including the use of variable steering wheel resistance to signal bus operators that they are departing from a designated bus lane.
Unintended consequences
Today, a number of in-vehicle technologies already exist that address drivers' perceptual and physical limitations. One example is electronic stability control, which improves vehicle controllability by allowing an onboard computer to control the braking of individual wheels and, in some cases, adjust engine power during sudden maneuvers. Onboard GPS systems that reduce the cognitive demands of navigation are proving to be popular optional equipment. New technologies, such as adaptive cruise control systems that automatically maintain a safe headway between vehicles on the highway are now entering the consumer market, while even more advanced vehicle systems incorporating vision enhancement and obstacle detection are on the horizon.
Because these systems make driving physically and cognitively easier, Manser says, they all have the potential to improve driver performance by freeing the driver's resources to focus on primary driving tasks. However, he cautions, technologies can have unintended consequences when they are used in ways their designers never intended. For example, instead of concentrating on driving, people may choose to use their extra cognitive resources to fine-tune the stereo system or chat on their cellular phones.
Manser first set out to determine whether drivers could effectively process information presented to them through this novel non-visual channel. The HumanFIRST Program's immersive driving simulator provided an ideal environment for initial testing, allowing the research team to monitor driver reactions and control the parameters of driving situations in which the haptic feedback system would be activated.
Reaction time data from initial tests revealed that the drivers responded to the sudden slowdown by a lead vehicle by moving their feet off the accelerator pedal more quickly when using the haptic feedback system, and that this benefit was present in both high-complexity and low-complexity secondary task scenarios.
A second finding from this test highlights some of the hidden complexity of driver response. The data reveal that the initial reaction times of drivers using the haptic feedback system are better (i.e., lower), but that the drivers then take slightly longer to transition from the accelerator pedal to the brake pedal. Manser believes this slight delay may be the result of drivers performing a visual double-check on the lead vehicle to make sure it is actually slowing. This extended transition time is more than offset by the reduced reaction time, making total response time significantly better with haptic feedback than without.
These results substantiate the hypothesis that the use of a haptic feedback system can result in a significant improvement in driver performance, and that this is the result of the system freeing cognitive resources that are then directed to the primary task of driving. However, an unintended consequence is that drivers could use the cognitive resources it frees up to perform other secondary tasks.
The HumanFIRST researchers devised a second test to investigate this possibility. Like the first test, it involved performing a secondary task while following a lead vehicle, but this time the task required drivers to interact with a touch-screen display rather than adjusting a stereo system. This task was designed to be more demanding perceptually, cognitively, and physically in order to approximate the normal demands of an in-vehicle secondary task (i.e., cellular phone) of real-world situations. Drivers were instructed to complete the task as many times as they could within a two-minute period.
The HumanFIRST Program’s virtual driving
simulator.
In the second set of tests, the researchers found that using the haptic feedback system improved vehicle controllability (the primary task), but also improved drivers' performance on the secondary task. These findings suggest that drivers are taking advantage of the newly freed resources from the haptic system to improve driving performance and to improve secondary task productivity. In 2006, HumanFIRST received follow-on funding from Nissan to continue researching the implications of the haptic pedal system.
The future of feedback
Haptic feedback is one example of how technology can be used to support and enhance the abilities of drivers. As researchers and automotive manufacturers continue to work together to develop new driver support systems, the future may see further improvements in driving performance.
"Today, we are beginning to develop technologies that address human performance limitations in completely new ways. These tools have the potential to improve performance and make driving easier and more comfortable—but, we need to understand first and foremost how human beings interact with new technologies."
— Peter Park Nelson