Research Papers

Task Analysis of Manual Phone-Dialing in Nine Production Vehicles

Filename 7A_3_Lavalliere_Paper.pdf
Filesize 635 KB
Version 1
Date added June 28, 2017
Downloaded 1 time/fois
Category 2017 CARSP XXVII Toronto
Tags Research and Evaluation, Session 7A
Author/Auteur Martin Lavallière, Bruce Mehler,  Joonbum Lee, Linda Angell, Bobbie Seppelt, Bryan Reimer
Stream/Volet Research and Evaluation
Slidedeck Presentation

7A-3_Lavallière et al Cellphone workflow


The use of cell phones while driving has become increasingly commonplace. As original equipment manufacturers race to implement integrated cellular technology into the vehicle, a growing body of research suggests links between increased use of technology and unsafe driving. Since the design choices in these Human Machine Interfaces (HMIs) such as the use of lists, hierarchical menus and multi-touch controls likely impact both the time and cognitive load associated with their use, we set out to first catalog a cross-section of visual-manual cell phone implementations in current production vehicles. Since research suggests specific design choices such as number and type of interactions can affect driver behavior in completing a task (Mizobuchi et al., 2013; Salvucci, 2001), the specific sequence of interactions in a DVI may potentially be important for driver safety. In this study, analysis focused on cell phone implementations made available through production vehicle DVI. Nine vehicles' infotainment systems were tested and compared. Following a methodology outlined by Reagan and Kidd (2013), we utilized hierarchical task structure method to examine current implementations of a standardized, manual cell phone dialing task. Through the use of hierarchical task analysis, we found differences in the number and types of interactions currently employed. On average, dialing a phone call took (M=5.17, SD=1.37) operations. The easy and hard dialing took an average of 4.56 and 5.78 operations to be completed, respectively. By region, the American model required (M=6.0, SD=1.79); the European models (M=4.00, SD=1.10); and the Asian models (M=5.50, SD=1.22). Overall, the results of our task analysis found a good deal of variability across manufacturers, regions, and task. While in general, the "hard" task required more steps than the "easy", in fact, two models evaluated in the present study required the same number of operations for both. These data suggest future avenues of research on managing a driver's visual, manipulative and cognitive demand, and the continued importance of human factors research in automotive interface development.

Martin Lavallière, Bruce Mehler,  Joonbum Lee, Linda Angell, Bobbie Seppelt, Bryan Reimer