Juan Gilbert, Aqueasha Martin, Wanda Eugene, Hanan Alnizami, Wanda Moses, Deidra Morrison
Over the past few years, the issue of distracted driving has gained the attention of the public and of policy makers. As a safety concern, distracted driving is an international issue. Tragic accidents attributed to texting while driving have flooded the media, leaving in their wake families and friends devastated by the consequences. Recently, in Anderson, S.C., a car accident occurred when the driver of a minivan suffered a head-on collision with another car. After the accident, the driver of the offending car admitted that she was texting and crossed the center line. The driver of the minivan died at the scene.
A study by Virginia Tech's Transportation Institute found that operators of commercial vehicles participating in texting while driving were more than 20 times more likely to become involved in a safety-critical event. Another study shows that 80 percent of all crashes and 65 percent of near crashes are due to a driver not paying attention. Additionally, a study conducted at Clemson University by psychology professor Johnell Brooks found that drivers using mp3 devices, such as iPods, or participating in texting while driving were 10 percent more likely to swerve from their lanes. The results of this study were later presented at the January 2008 Annual Meeting of the Transportation Research Board in Washington, D.C. Policy makers have taken notice and put into action steps to eliminate the number of accidents resulting from distracted driving.
In 2009 President Obama signed an Executive Order that banned texting for federal employees while driving government-owned vehicles or with government-owned equipment. In 2010 U.S. Transportation Secretary Ray LaHood announced a federal ban on texting by drivers of commercial vehicles, such as large trucks and buses. Currently, 19 states, the District of Columbia, and Guam have passed laws that ban all drivers from texting while driving. However, nearly all 50 states have some type of partial ban on the use of handheld devices while driving. In addition at least 52 countries have bans in some or all areas. Fines are the major penalty for most of these countries, but in some countries, such as Britain, texting while driving is also punishable by a jail sentence. Despite these bans, the numbers of accidents have not decreased. Although the dangers of distracted driving are apparent, the practice remains a problem. The reality is people break laws, and despite the known dangers and risks, they will continue to text while driving. This article discusses the text-driving epidemic and the societal response and introduces an alternative approach called voiceTEXT.
The issue of distracted drivers can be tied to recent research that shows attention sharing generated by phone use while at the wheel appears to increase the driver's mental workload, thereby overloading the driver's cognitive capacities and impairing driving performance . Behaviors such as impaired gap judgment, reduced sensitivity to road conditions, poor lane maintenance, and the increase in reaction times to driving-related events (e.g., brake lights, etc.) can all be tied to distracted drivers. Straye et al. found these behaviors as profound as those associated with driving while drunk. In addition, research suggests that when using secondary in-car devices such as cell phones, the risks associated with driving increase considerably as the amount of interaction with the device required of the driver increases .
Newer generations of phones have adopted touch-screen technology. With such technology and with the absence of physical buttons, constant visual scanning of the screen is necessary to navigate a phone application. Phones not equipped with touch screens often include a physical keypad that is used to help navigate different tasks on a mobile phone. Previous research suggests this type of tactile feedback is important to measure usability, performance, and user experience. With physical keys a user is often able to learn how to use the device without having a visual scan. It was found that a novice user is able to improve their learning curve after using a typical Dual-Tone Multi-Frequency (DTMF) keypad for calling or texting many times. A user can remember center keys such as "2," "4," "6," "8," and/or "0." If a user was texting with this layout, it was found that he or she keeps count of how many times a button is pressed to reach the desired letter. Such tactile feedback is necessary once a user obtains a visual memory of the layout. Touch screens, on the other hand, are not equipped with physical feedback. Several cutting-edge mobile phones have large touch screens that are unequipped with physical buttons. Instead, other verification responses such as vibrations or sound are used. These responses, however, verify only that a button has been pressed, not that the correct button has been pressed; therefore, visual scanning of the screen is necessary to ensure correct navigationincreasing a driver's distractions.
A report on commercial vehicles found that texting increased the likelihood of a safety-critical event more than 23 times, followed by other complex tasks such as rummaging through a grocery bag (See Table 1). In addition, studies have been conducted to gauge the effects of hands-free devices on cognitive load while driving. One study found that the use of a hands-free device was similar to that of having a conversation with a passenger; however, this was dependent on, among other factors, the complexity of the conversation (low demanding to high demanding).
Many of the distractions in Table 1 have been around for some time, and there are few policies that specifically cover many of them. Like texting, many are seldom noticed unless one of them leads to some other infraction. Additionally, it's difficult for law-enforcement officers to notice many of these behaviors. The findings on texting while driving, however, have directed governments into action to enact policies. But research shows they have not had as much of an impact on texters. Bruyas et al. suggests that banning legislation has a substantial short-term effect that wears off over time . A driver study showed that 84 percent of drivers believe they drive in a less safe manner when engaged in other tasks. Yet 71 percent of people between the ages of 18 and 49 admit they text or talk on the phone while they drive. Texting while driving has become a dangerous habit that is widespread and socially acceptable. Although bans on texting while driving have been implemented, catching someone texting and driving is not an easy task.
Laws being passed intend to discourage texting while driving, but these laws may prove to be a major challenge to enforce. Texting is often an unseen activity concealed by the size of the cellular device and privacy mechanisms, such as window tint. In some cases law-enforcement officials must have other forms of just cause, such as speeding, before issuing a citation for texting. And finding ways to catch perpetrators prior to an accident may not be as easy as catching those who violate other moving-vehicle laws (i.e., speeding, not wearing seatbelts, DUI, running a red light, etc.). Current methods of enforcing moving-vehicle laws were not designed to catch a texter.
Radar detectors are useful for catching drivers who break speed-limit laws but are limited in their ability to catch someone texting. Similarly, traffic cameras placed at strategic locations are useful for identifying someone not wearing a seat belt or someone who runs a red light, but this method would not be as effective for identifying someone texting while driving: Texting is sporadic and may happen at other times.
Additionally, people text with their phone in different positions: An upward texter, someone who positions the phone above the steering wheel, is easier to catch than a downward texter, someone who holds the phone below the dashboard. Direct observation allows law-enforcement officials to identify persons driving under the influence, although other tests are usually conducted once the driver has been stopped. This method may be the most viable for catching someone texting while driving, but it can be ineffective given that texting is not as easily detected. For example, how do you determine the driver was texting and not dialing a phone number? Besides, if law-enforcement officers were to seize a driver's phone, this could violate the individual driver's privacy given the personal information contained on a phonephotos, contacts, etc. Other methods for catching someone texting while driving, such as a volunteered statement from the driver, or accessing phone records, would likely transpire after an accident has occurred, and in addition would require more time.
So, what are the alternatives? One of the major challenges in developing in-vehicle systems as this market continues to expand is to design them in a way that will not take a driver's attention from the primary task of driving . Therefore, one very apparent alternative is to convince people to abstain from unsafe behavior. There is, however, no clear answer for why people text while driving even though the dangers are known, so laws that ban texting while driving will not completely eliminate the activity, a claim that research supports. Ultimately, there will be people who continue to text behind the wheel.
Although the dangers of distracted driving are apparent, the practice remains a problem. The reality is people break laws, and despite the known dangers and risks, they will continue to text while driving.
One common approach is to prevent a phone from being able to text while a vehicle is in motion. There are several variations of software available that either block cellular phone signals or block calls and messages when a vehicle's speed is above a certain point. Other such software may lock the cell phone's keys, allowing only emergency calls. Aimed at parents with teens, these software applications allow parents to control their teen's texting behaviors when driving. These implementations are not met without controversy. If you are a passenger in a vehicle, these applications will disable your phone as well. Then there is the issue of emergency calls. If a parent gets an emergency call while driving, some of these applications would send the call directly to voicemail.
With the rise of mobile phone applications, voice applications such as voice memos and voice dialing are becoming standard on mobile phones. Using the Apple iPhone's voice-control features, a user can make phone calls, play music, and control GPS navigation systems using their voice . In addition, other companies have the means to perform voice-activated searches. ShoutOut, Dragon, and VLingo offer speech as a means of sending text and email messages. From a policy viewpoint, such features are praised for offering alternatives to drivers. However, there are caveats: These features can still lead to cognitive overload and provide additional distractions as drivers engage in continuous dialog with their systems. In addition, a novice user may seek visual confirmation of their request.
Even so, these voice-activated applications exist as a viable alternative. There are several applications available that read a recipient's text message aloud, while others allow voice-activated text conversations. Some applications also transcribe spoken text messages. Voice-to-text applications installed locally on your phone typically transcribe your speech into text, give you the option to review the text, and then send it as SMS text or email. Given that speech recognition for transcribing text isn't 100 percent accurate, users tend to look at the text and read it before sending it. They want to make sure the voice-to-text application gets the transcription correct; otherwise the user could send an incorrectand potentially offensivemessage. The fact that the user or driver can not completely trust speech recognition for transcription ultimately forces the user to look at the phone, defeating the purpose of going hands-free. Public policy is likely to look favorably upon voice-to-text applications if they can be used in a hands-free, eyes-free mode. Additionally, we recommend that anytime a text is transcribed and sent via SMS or email, the message should include a disclaimer that the text may be incorrect or incomplete due to inaccuracy in speech recognition. In testimony before the U.S. House of Representatives Committee on Transportation and Infrastructure, Thomas A. Dingus, director of Virginia Tech's Transportation Institute, suggested a well-designed "true hands-free" device, such as voice-activated systems, may be the answer to the issue of texting while driving.
voiceTEXT is a tool that sends voice messages instantly over a phone line. voiceText was implemented using VoiceXML (Voice eXtensible Markup Language) and CCXML (Call Control XML) by researchers in the Human Centered Computing Lab at Clemson University. It is a voice-activated system designed and developed to offer a safe, hands-free, eyes-free alternative to texting while driving. It allows people to stay connected, while eliminating the need for a driver to take his or her eyes off the road. The process begins when the sender establishes a connection with the server using a phone (cellular, land-line, Internet, e.g., voice over IP). After the connection has been established, the sender is prompted to compose a message, which is recorded and saved on the server. The sender will verbally specify the recipient by name, and the voiceTEXT system then places a phone call to the recipient. When the recipient answers the call, the voiceTEXT message is played to them and the recipient is prompted to perform an action. The recipient may end the call using his or her voice, replay the message, forward the message, or reply to the message. Using this system, two drivers can voiceTEXT each other without ever taking their eyes off the road.
In addition, voiceTEXT provides SMS texting and email services. The recipient of a voiceTEXT will receive a voiceTEXT message as well as an email and SMS text. The body of the email or SMS text will contain one or more of the following:
- Phone number where the recipient can retrieve a voiceTEXT,
- Hyperlink to an audio file of the voiceTEXT that can be played from the phone or browser,
- Attachment of the voiceTEXT audio file,
- Transcribed text; that is, the sender's message transcribed using speech-to-text technology.
These options provide for a broader expansion of the concept of texting services while differing from other instant-messaging techniques in that:
- They use call connections or voice-over-IP to deliver the message versus SMS texting.
- The primary message is an audio stream versus text.
Because voiceText was designed to be an alternative to texting, it is different from voicemail or participating in a voice call with a hands-free cell phone. Voicemail, in which a user must call the recipient, wait until the voicemail prompt plays, and leave a message, was not designed for the immediate exchange of short voice messages. The recipient of a voicemail must access the voicemail system by calling a number or by some other means, search for, and play the message. With voiceText the sender's message is recorded and sent via a phone call to the receiver, not requiring any intermediate access on the receiver's part.
As mentioned earlier the complexity of the conversation has some effect on a driver's cognitive load. Whereas a hands-free phone conversation allows participants to interact continuously, voiceText allows only for short messages just as traditional texting. In this case the interactions are different. In a hands-free phone conversation, the rules of conversation apply: Each participant is expected to listen, understand, respond appropriately, respond in a timely fashion, pay attention, etc. However, the conversational rules of texting differ in that texting requires only that the recipient of the text read and respond when possible. voiceText was designed specifically for these interactions, not a full conversation.
Distracted driving is not a new issue, but as the number and variety of communication devices increase, so will the number of potential distractions. One key to mitigating these distractions will be to design technologies for safe driver interactions. Although bans may discourage some, others will continue to text and drive despite the dangers. Current research indicates that the number of accidents has not decreased. Previous research suggests asynchronous messaging like voiceTEXT results in less distraction than having a phone conversation. Additional studies are planned for voiceTEXT at Clemson University as well. The results of these studies will have the potential to provide additional information to lawmakers on texting while driving and provide direction for executing future policies.
This material is based upon work supported by the National Science Foundation under Grant # 0937060 to the Computing Research Association for the CIFellows Project. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or the Computing Research Association.
1. Apple Inc; http://www.apple.com/iphone/iphone-3gs/
2. Bach, K.M., Jaeger, M.G., Skov, M.B., and Thomassen, N.G. "Interacting with In-Vehicle Systems: Understanding, Measuring, and Evaluating Attention." Proceedings of the 2009 British Computer Society Conference on Human-Computer Interaction, 453462.
3. Box, S. VirginiaTech Transportation Institute. "New Data from VTTI Provides Insight into Cell Phone Use and Driving Distraction." July 27, 2009. http://www.vtti.vt.edu/PDF/7-22-09-VTTI-Press_Release_Cell_phones_and_Driver_Distraction.pdf/
4. Brumby, D.P., Howes, A., and Salvucci, D.D. "A Cognitive Constraint Model of Dual-Task Trade-offs in a Highly Dynamic Driving Task. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 233242, 2007. http://delivery.acm.org.proxy.lib.clemson.edu/10.1145/1250000/1240664/p233-brumby.pdf?key1=1240664&key2=9335089621&coll=ACM&dl=ACM&CFID=81873461&CFTOKEN=58472802/
5. Bruyas, M.P., Brusque, C., Tattegrain, H., Auriault, A., Aillerie, I. and Duraz, M. "Consistency and Sensitivity of Lane Change Test According to Driving Simulator Characteristics." IET Intelligent Transport Systems Journal 2, 4 (2008): 306314.
Juan E. Gilbert is a professor and chair of the Human Centered Computing Division in the School of Computing at Clemson University. His research interests are in spoken language systems, accessibility, multimodal user interfaces, culturally relevant computing, and science and technology policy.
Aqueasha M. Martin is a Ph.D. student in the Human Centered Computing Division in the School of Computing at Clemson University. She received a B.S. in computer science from Tougaloo College and a M.S. in computer science from the University of Alabama, Birmingham. Her research interests are in the areas of natural language processing, accessibility, and health informatics.
Wanda Eugene is a doctoral candidate in the Human Centered Computing Lab in the Computer Science and Software Engineering Department at Auburn University. She is interested in how cultural, social, and personal surroundings affect the appropriation of computational artifacts and ideas and how they can serve as a resource for the design of new technologies. Eugene received a bachelor's in electrical engineering from Florida State University, a master's in industrial engineering from the Florida Agricultural and Mechanical University, and a master's in interdisciplinary studies specializing in instructional technology and African-American studies from George Mason University.
Hanan Alnizami is a Ph.D. student and a research assistant in the Human Centered Computing Division in the School of Computing at Clemson University. She earned a bachelor's in information technology at Youngstown State University. Her research interests include spoken language and multimodal user interfaces.
Wanda Moses is a Ph.D. student in the Human Centered Computing Lab in the School of Computing at Clemson University. Her research interest is in designing culturally relevant educational tools to assist in the motivation, development and advancement of African-American students in the STEM disciplines. She received her master's degree from Auburn University in computer science and software engineering. Moses earned her B.S. in both mathematics and computer science at South Carolina State University. All of this was accomplished after retiring from 20 years of military service in the U.S. Navy. She is also a GEM graduate fellow.
Deidra Morrison is a post-doctoral researcher in the division of Human Centered Computing in the School of Computing at Clemson University. She received her B.S. in computer science from Spelman College, and her M.S. and Ph.D. from Northwestern University. She was a GEM graduate fellow, and was awarded the Computer Innovations Fellowship through the Computer Research Association and NSF. Morrison's area of research involves social computing and investigating social interaction in non-traditional spaces and relationship management using technology.
For complete reference list, visit our website: http://www.voice-textingresearch.org/DrivingPolicy.pdf
©2010 ACM 1072-5220/10/0700 $10.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2010 ACM, Inc.