Speech warnings: a review

The effects of various auditory takeover requests: A simulated driving study considering the modality of non-driving-related tasks

With the era of automated driving approaching, designing an effective auditory takeover request (TOR) is critical to ensure automated driving safety. The present study investigated the effects of speech-based (speech and spearcon) and non-speech-based (earcon and auditory icon) TORs on takeover performance and subjective preferences. The potential impact of the non-driving-related task (NDRT) modality on auditory TORs was considered. Thirty-two participants were recruited in the present study and assigned to two groups, with one group performing the visual N-back task and another performing the auditory N-back task during automated driving. They were required to complete four simulated driving blocks corresponding to four auditory TOR types. The earcon TOR was found to be the most suitable for alerting drivers to return to the control loop because of its advantageous takeover time, lane change time, and minimum time to collision. Although participants preferred the speech TOR, it led to relatively poor takeover performance. In addition, the auditory NDRT was found to have a detrimental impact on auditory TORs. When drivers were engaged in the auditory NDRT, the takeover time and lane change time advantages of earcon TORs no longer existed. These findings highlight the importance of considering the influence of auditory NDRTs when designing an auditory takeover interface. The present study also has some practical implications for researchers and designers when designing an auditory takeover system in automated vehicles.

Test Pilot and Airline Pilot Differences in Facing Unexpected Events

BACKGROUND: Unexpected events in flight might decrease the transparency of the flying process and weaken the pilot's perception of the current state, or even erode manipulating skills. However, during the flight test of a new or modified aircraft, to verify the boundaries of aircraft aerodynamic performance and handling stability, unexpected events may be encountered that need to be handled by the test pilot. Therefore, studying the differences between test pilots and airline pilots could help improve flight safety.METHODS: Two kinds of physiological parameters, eye blink rate and average fixation duration and task-related performance of test pilots and airline pilots, were analyzed in three abnormal scenarios. A total of 16 pilots participated. The study was carried out in an A320 flight simulator.RESULTS: The differences were significant for both test pilots and airline pilots in eye blink rate and average fixation duration. Furthermore, the reaction time of test pilots (Mean = 23.38 s) was significantly shorter than airline pilots (Mean = 42.63 s) in Unreliable Airspeed condition, and the pitch angle deviations between them were significant in both Wind Shear and Unreliable Airspeed condition.DISCUSSION: The uncertainty of environmental change could create more severe pressure and mental workload influence than actual system failure. For airline pilots, compared with test pilots, the importance of practicing manual flight should still be emphasized. Improving reactions to unexpected ambient conditions and unannounced fault status could also contribute to flight safety.Zheng Y, Lu Y, Jie Y, Zhao Z, Fu S. Test pilot and airline pilot differences in facing unexpected events. Aerosp Med Hum Perform. 2023; 94(1):18-24.

Novel sonification designs: Compressed, iconic, and pitch-dynamic auditory icons boost driving behavior

With the development of connected vehicles, in-vehicle auditory alerts enable drivers to effectively avoid hazards by quickly presenting critical information in advance. Auditory icons can be understood quickly, evoking a better user experience. However, as collision warnings, the design and application of auditory icons still need further exploration. Thus, this study aims to investigate the effects of internal semantic mapping and external acoustic characteristics (compression and dynamics design) on driver performance and subjective experience. Thirty-two participants (17 females) experienced 15 types of warnings - (3 dynamics: mapping 0 vs. 1 vs. 2) × (5 warning types: original iconic vs. original metaphorical vs. compressed iconic vs. compressed metaphorical auditory icon vs. earcon) - in a simulator. We found that compression design was effective for rapid risk avoidance, which was more effective in iconic and highly pitch-dynamic sounds. This study provides additional ideas and principles for the design of auditory icon warnings.

Speech-based takeover requests in conditionally automated driving: Effects of different voices on the driver takeover performance

This study explored the possibility of applying personalized takeover requests (TORs) in an automated driving system (ADS), which required drivers to regain control when the system reached its limits. A driving simulator experiment was conducted to investigate how speech-based TOR voices impacted driver performance in takeover scenarios with two lead time conditions in conditionally automated driving (level 3). Eighteen participants drove in three sessions, with each session having a different TOR voice (a synthesized male voice, a synthesized female voice, and a significant other voice). Two scenarios with a lead time of 5 s and two scenarios with a lead time of 12 s were provided per session. The driver takeover time and quality data were collected. A follow-up interview was conducted to gain a clearer understanding of the drivers' psychological feelings about each TOR voice during takeovers. Changes in takeover time and takeover quality caused by TOR voices were similar in both lead time conditions, except for the lateral acceleration. The synthesized male voice led to a larger maximum lateral acceleration than the other two voices in the 5 s condition. Interestingly, most drivers preferred choosing the synthesized female voice for future takeovers and showed negative attitudes toward the significant other voice. Our results implied that choosing TOR voices should consider the drivers' daily voice-usage habits as well as specific context of use, and personalized TOR voices should be incorporated into the ADS prudently.

Mitigating Patient and Consumer Safety Risks When Using Conversational Assistants for Medical Information: Exploratory Mixed Methods Experiment

BACKGROUND

Prior studies have demonstrated the safety risks when patients and consumers use conversational assistants such as Apple's Siri and Amazon's Alexa for obtaining medical information.

OBJECTIVE

The aim of this study is to evaluate two approaches to reducing the likelihood that patients or consumers will act on the potentially harmful medical information they receive from conversational assistants.

METHODS

Participants were given medical problems to pose to conversational assistants that had been previously demonstrated to result in potentially harmful recommendations. Each conversational assistant's response was randomly varied to include either a correct or incorrect paraphrase of the query or a disclaimer message-or not-telling the participants that they should not act on the advice without first talking to a physician. The participants were then asked what actions they would take based on their interaction, along with the likelihood of taking the action. The reported actions were recorded and analyzed, and the participants were interviewed at the end of each interaction.

RESULTS

A total of 32 participants completed the study, each interacting with 4 conversational assistants. The participants were on average aged 42.44 (SD 14.08) years, 53% (17/32) were women, and 66% (21/32) were college educated. Those participants who heard a correct paraphrase of their query were significantly more likely to state that they would follow the medical advice provided by the conversational assistant (χ²₁=3.1; P=.04). Those participants who heard a disclaimer message were significantly more likely to say that they would contact a physician or health professional before acting on the medical advice received (χ²₁=43.5; P=.001).

CONCLUSIONS

Designers of conversational systems should consider incorporating both disclaimers and feedback on query understanding in response to user queries for medical advice. Unconstrained natural language input should not be used in systems designed specifically to provide medical advice.

Analyzing crowdsourced ratings of speech-based take-over requests for automated driving

Take-over requests in automated driving should fit the urgency of the traffic situation. The robustness of various published research findings on the valuations of speech-based warning messages is unclear. This research aimed to establish how people value speech-based take-over requests as a function of speech rate, background noise, spoken phrase, and speaker's gender and emotional tone. By means of crowdsourcing, 2669 participants from 95 countries listened to a random 10 out of 140 take-over requests, and rated each take-over request on urgency, commandingness, pleasantness, and ease of understanding. Our results replicate several published findings, in particular that an increase in speech rate results in a monotonic increase of perceived urgency. The female voice was easier to understand than a male voice when there was a high level of background noise, a finding that contradicts the literature. Moreover, a take-over request spoken with Indian accent was found to be easier to understand by participants from India than by participants from other countries. Our results replicate effects in the literature regarding speech-based warnings, and shed new light on effects of background noise, gender, and nationality. The results may have implications for the selection of appropriate take-over requests in automated driving. Additionally, our study demonstrates the promise of crowdsourcing for testing human factors and ergonomics theories with large sample sizes.

Effect of directional speech warnings on road hazard detection

BACKGROUND

In the last 2 decades, cognitive science and the transportation psychology field have dedicated a lot of effort to designing advanced driver support systems. Verbal warning systems are increasingly being implemented in modern automobiles in an effort to increase road safety.

OBJECTIVE

The study presented here investigated the impact of directional speech alert messages on the participants' speed to judge whether or not naturalistic road scenes depicted a situation of impending danger.

METHOD

Thirty-eight volunteers performed a computer-based key-press reaction time task.

RESULTS

Findings indicated that semantic content of verbal warning signals can be used for increasing driving safety and improving hazard detection. Furthermore, the classical result regarding signal accuracy is confirmed: directional informative speech messages lead to faster hazard detection compared to drivers who received a high rate of false alarms.

CONCLUSION

Notwithstanding some study limitations (lack of driver experience and low ecological validity), this evidence could provide important information for the specification of future Human-Machine-interaction (HMI) design guidelines.