United States Department of Transportation
i
In-Vehicle Voice Control Interface Evaluation: Preliminary Driver Workload and Risk Analysis
-
2020-01-01
-
[PDF-6.51 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
DOI:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report
-
Contracting Officer:
-
Corporate Publisher:
-
Abstract:This project evaluated distraction and relative risk associated with using voice control systems (VCS) while driving. The objective was to explore potential empirical measures and to use a modeling approach for evaluating risk with these voice-based systems. The project included three studies. Study 1 and Study 2 were designed to assess potential measures of the workload and demands on the driver imposed by voice-based and hybrid (audio plus visual) tasks. Participants interacted with a “Wizard of Oz” VCS while driving and a novel radio tuning benchmark task was used. Study 1 (n = 9) compared response times for the tactile detection response task (TDRT), which is a standardized measure of cognitive load, with a modified remote detection response task (RDRT) which was implemented in this study to provide a complementary measure of visual attention toward the forward roadway. Response time was sensitive to differences between VCS tasks, and there was a significant interaction between detection response task (DRT) type and VCS task. Study 2 (n = 9) included both on-road data collection and data collection with a driving simulator. In three different sessions, measurement protocols included TDRT, response time to the lead vehicle (LV) brake light, speed matching to a speed-varying LV, and off-road eye glance measures. Results indicated that task completion time, TDRT performance, and eye glance measures distinguished between VCS tasks and provided similar results in the on-road and simulated driving contexts. Study 3 used an analytical approach to develop relative risk estimates and crash severity estimates for the VCS tasks tested in this project. Counterfactual -- or “what if” -- simulations made use of a set of crash events and near crash events recorded in the second Strategic Highway Research Program (SHRP 2) naturalistic driving study. These events were then replayed as a set of simulations, where the eye glance data collected in Study 1 and Study 2 were substituted for the original eye glance data recorded from drivers in the SHRP 2 study. Risk estimates and crash severity estimates developed using this technique varied considerably by VCS task and by driver.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
