The smartphone and the driver’s cognitive workload: A comparison of Apple, Google, and Microsoft’s intelligent personal assistants.

Abstract

The goal of this research was to examine the impact of voice-based interactions using 3 different intelligent personal assistants (Apple's , Google's for Android phones, and Microsoft's ) on the cognitive workload of the driver. In 2 experiments using an instrumented vehicle on suburban roadways, we measured the cognitive workload of drivers when they used the voice-based features of each smartphone to place a call, select music, or send text messages. Cognitive workload was derived from primary task performance through video analysis, secondary-task performance using the Detection Response Task (DRT), and subjective mental workload. We found that workload was significantly higher than that measured in the single-task drive. There were also systematic differences between the smartphones: The Google system placed lower cognitive demands on the driver than the Apple and Microsoft systems, which did not differ. Video analysis revealed that the difference in mental workload between the smartphones was associated with the number of system errors, the time to complete an action, and the complexity and intuitiveness of the devices. Finally, surprisingly high levels of cognitive workload were observed when drivers were interacting with the devices: "on-task" workload measures did not systematically differ from that associated with a mentally demanding Operation Span (OSPAN) task. The analysis also found residual costs associated using each of the smartphones that took a significant time to dissipate. The data suggest that caution is warranted in the use of smartphone voice-based technology in the vehicle because of the high levels of cognitive workload associated with these interactions. (PsycINFO Database Record

Talking to your car can drive you to distraction

“…In fact, many newer-model vehicles come equipped with voice-activated systems that allow drivers to adjust climate control, select music, place and receive phone calls, send and read textual messages, and interact with social media. The impact of these voice-based commands on the motorist’s cognitive workload is unknown; however, voice-command laboratory surrogates with perfect reliability have been shown to produce surprisingly high levels of workload (Strayer, Cooper, Turrill, Coleman, & Hopman, 2015a, Strayer et al, 2015b). Moreover, older adults are likely to have greater difficulty using these voice-based systems and, ironically, they are more likely to purchase new vehicles equipped with these features (Sivak, 2013).…”

“…Our prior research provided a benchmark for the cognitive workload associated with common in-vehicle activities (Strayer et al, 2015a, 2015b; see also Cooper, Ingebretsen, & Strayer, 2014; and Strayer, Turrill, Coleman, Ortiz, & Cooper, 2014). In our studies, we developed and validated a cognitive distraction scale based on converging operations from the laboratory, driving simulator, and an instrumented vehicle driven in a residential section of Salt Lake City.…”

“…The primary difference between our laboratory-based speech-to-text system and the Siri -based interactions was the reliability of the system (see also Strayer et al, 2015a). Siri was error-prone, producing different responses to seemingly identical commands.…”

Multi-modal demands of a smartphone used to place calls and enter addresses during highway driving relative to two embedded systems

“…Concerns have been raised about the cognitive demands of tasks that still remain with voice interfaces (Cooper, Ingebretsen, and Strayer 2014; Reimer, Mehler, McAnulty, et al 2013; Reimer et al 2010, 2012; Strayer et al 2013; Strayer 2015a, 2015b; Strayer et al 2014). At the same time, several studies have found that self-reported workload, physiological arousal (e.g.…”

Multimodal speech and gesture control of AGVs, including EEG-based measurements of cognitive workload

“…Combined speech and gesture control is the current state of the art in many applications. Whereas gesture control primarily requires physical effort, speech control involves predominantly cognitive effort, as previously demonstrated with regard to the speech control of mobile phones during driving tasks [2]. In 2013, more than ten thousand industrial accidents involving fork lift trucks were reported in Germany, approximately one-third with fatal consequences [3].…”