Predicting the Effects of Driver Distraction by On-Board Devices: An Integrated Model Approach

Speaker
Dario Salvucci
Post-Doctoral Research Associate, Cambridge Basic Research, Nissan Technical Center North America

When
-

Where
Newell-Simon Hall 1305 (Michael Mauldin Auditorium)

Description

As drivers gain access to increasingly numerous and complex on-board devices for support (e.g., navigation) and “infotainment” (e.g., news and e-mail), many have raised serious concerns about the issue of driver distraction—excessive attention to secondary devices rather than the primary driving task. In this talk I will discuss how we can better understand and alleviate driver distraction by modeling driver behavior in a cognitive architecture. A cognitive architecture is a computational framework that incorporates built-in, well-tested parameters and constraints for cognitive and perceptual-motor processes. We have developed a driver model in the ACT-R architecture that combines lower-level controllers for speed and lateral position with a higher-level cognitive process for situation awareness and task management. When integrated with models of secondary-task behavior, the driver model generates a priori behavioral predictions about how secondary tasks affect driver performance. Such predictions can in turn be used to evaluate and compare on-board devices to facilitate development and testing. To illustrate this approach, I will describe an integrated model of cell-phone dialing and driving and validate the model’s predictions with control and eye-movement data collected from human drivers in a fixed-base driving simulator.

Speaker's Bio

Dario Salvucci is a postdoctoral researcher at Nissan Cambridge Basic Research in Cambridge, MA. He received his Ph.D. in computer science in 1999 at Carnegie Mellon, where his dissertation investigated automated methods for interpreting human eye movements. His current research explores computational methods of generating and interpreting human cognition and behavior, most recently in the context of user interfaces and driving.

Host
John Anderson