SLIDE 1: Driver Distraction: Understanding the Problem, Identifying Solutions January 7, 2005 Joseph N. Kanianthra, Ph.D. Associate Administrator for Vehicle Safety Research National Highway Traffic Safety Administration SLIDE 2: What is Driver Distraction? Distraction occurs when the driver shifts attention away from the primary driving to a thought, event, activity, or person. Distracting sources have always been present New sources are increasing in numbers and complexity—both OEM technologies and aftermarket devices New sources can affect driver behavior and performance in different ways than conventional distractions We do not know what will be a source of driver distraction in the future as technology is advancing at a rapid rate, e.g., more voice interfaces SLIDE 3: High Technology vs Low Technology Distractions May engage attention longer and more frequently May place more cognitive and manual demands on drivers May interrupt drivers at unsafe times Interruptions can also be defined as anything that draws the attention of the driver. SLIDE 4: The Safety Problem of Electronic Distractors - Recognized by many manufacturers - Crash data not complete regarding existing sources of distraction Driver dishonesty, driver/witness misperceptions, low police investigation priority contribute to data limitations Problems from new “electronic distractors” may not show up in crash data for years Data clear that existing “electronic distractors” (primarily cell phones and radios) contribute to crashes, but extent of problem is unknown SLIDE 5: Distraction and Crash Risk: NHTSA Research Focus NTHSA research focuses on two area: Driver Willingness to Use and Distraction Demands. The remainder of the presentation will include research from both areas Crash risk results from the interaction of: How willing is the driver to perform a task while driving, The level of workload demand that the device use places on the driver, and The current demand of the driving scenario, which potentially places additional demands on the driver. SLIDE 6: Willingness to Engage While Driving NHTSA Research: Willingness to Engage Purpose: to collect ratings of willingness to engage, assessed risk for different tasks, and driving conditions Procedure: On-road experiment where participants gave ratings for their willingness to engage in different tasks under a range of driving scenarios. We asked for their subjective ratings under different conditions - they did not complete tasks Results: Willingness to engage correlated with drivers’ perceived crash risk The graphic depicts how participants rated risk for different in-vehicle tasks The data will help in formulating safety messages and in understanding how different drivers will react to different countermeasure approaches. For example, teen drivers were much more willing to multitask than older drivers. SLIDE 7: Inventory of Navigation Interface Designs: Task Demand Graphic Interpretation: For each of the entry type (Destination Entry), the minimum and maximum number of keystrokes is presented across navigation systems. The implication is that the fewer numbers of times a driver has to manually press a button, the better. This figure shows that there are large differences in navigation system interface designs that can affect how many “button presses” are needed to enter an address. The task shown requires a wide range of keystrokes. This emphasizes the importance of interface design. Note the differences in number of keystrokes is attributable to both the use of shortcuts (fewer keystrokes) and length of street name (more keystrokes). Project Description: We conducted an inventory of a sample of different in-vehicle technologies (nearly 80 technologies total) a few years ago. The inventory examined market-ready in-vehicle products (both OEM and aftermarket products), and identified a range of interface design features (e.g., control and displays characteristics, safety features, etc.) noting aspects and dimensions that have implications for potential driver distraction. The review was not intended to be exhaustive, but rather provide a diverse and representative range of system designs and configurations. SLIDE 8: How Interface Design Can Influence Driver Performance Graphic Interpretation: From this graphic, we can see that those tasks that allow drivers to keep their eyes on the road result in fewer lane exceedences (i.e., less driving degradation). That is why the voice interface results in no lane exceedences in this test. However, this is not an indication of cognitive distraction, and therefore does not present a complete picture of crash risk. We are continuing to research the effect of voice interfaces to help better understand the effect of cognitive demand on driving performance. SLIDE 9: 100-car Naturalistic Driving Study Goals: Understand the preceding factors associated with crashes, near crashes, critical events Develop relationship between task completion time, eyes-off-road time and critical incident likelihood Provide baseline relating performance to safety-related risk Overview: 1 year, 43K hours, 1.37M miles Approx. 76 crashes recorded, with about 38% related to driver distraction Will also be looking at near crashes Research questions include: Assessment of willingness to engage in and associated risk of distracting activities Types of critical events related to distraction Potential role of crash warning systems in preventing distraction related crashes SLIDE 10: 100-Car Naturalistic Driving Study Data Collection Capabilities NHTSA Research: 100-car Naturalistic Driving Study Five views are captured by cameras and recorded: driver’s face, driver’s lap, right rear, forward, and rear Video clip of participant talking on cell phone, and shows the capabilities of the data collection system. The findings will be released this year. SLIDE 11: CAMP - Driver Workload Metrics Project CAMP is a collaboration between NHTSA and Ford, GM, Nissan, and Toyota. The goal of CAMP is to develop: - Driver distraction metrics - Tests that are Practical, Meaningful, Repeatable - Human factors guidelines - Lab-based surrogate measures - And to explore a range of on-road performance measures SLIDE 12: Driver Assistance Systems To Alert Distracted Drivers These technologies are challenging because of the need to ensure correct detections of imminent crashes without giving drivers unacceptable false alarms. SLIDE 13: Adaptive Interface Workload Management SAfety VEhicle Using Adaptive Interface Technology We are exploring in this program the benefits that might be achieved if the driver state and driver intentions can be detected in real time to enhance the effectiveness of collision warning systems as well as to help minimize unsafe multitasking. Delphi Delco is working with us to explore the feasibility and benefits of this concept. Sensors outside the vehicle record driving task workload levels. Sensors inside the vehicle record driver distraction, possibly including direction of gaze. Algorithms determine in real time any mismatches between primary task demands and driver level of attention. Adaptive interface possibilities may include locking out information or changing the threshold for triggering crash warnings. The goal is to explore ways to monitor distraction in real time and automatically adapt the interface depending on the driving task demands and level of driver distraction away from those tasks. The research is developing a concept vehicle to evaluate the potential of this countermeasure concept. It is scheduled to be completed in 2006. SLIDE 14: In conclusion… Closing remarks Driving is a complex task, requires a high level of demand Need for driver awareness?