Human - System Cooperation in Automated Driving

Abstract

This special issue covers numerous challenges of growing relevance in the domain of human-system cooperation in automated driving. Nine articles out of 21 submissions were accepted for publication after two rounds of rigorous review. Throughout this issue, we share valuable findings based on case study, real world data, experiment, survey and review. The first two articles set high level understandings of automated driving and human-system cooperation. Liu, Yang, Wang, and Liu (Evaluating Initial Public Acceptance of Highly and Fully Autonomous Vehicles [AV]) performed a between-subject survey (N = 742) to study public perceptions and acceptance of highly and fullyAV. A psychological model, based on trust, was developed to explain three acceptance measures. The respondents showed a stronger belief of benefits from Fully AV than from Highly AV. Trust in AV retained a direct and an indirect effect (mainly through perceived benefit) on the three acceptance measures. Perceived benefit, in comparison to perceived risk, showed a higher direct effect on AV acceptance and a higher mediating effect on the trust-acceptance relationship. A prediction analysis demonstrated the model’s acceptable predictive capability for public acceptance. Biondi, Alvarez, and Jeong (Human-Vehicle Cooperation in Automated Driving: A Multidisciplinary Review and Appraisal) provided insights into how human drivers and vehicle systems interplay and influence each other. The article discusses: limitations of technology-centered automation taxonomies; benefits of accounting for human agents; contributions of machine learning to automated driving and how critical models in human-system cooperation can inform the design of a more symbiotic relationship between driver and vehicle; challenges in the human element to enable the safe introduction of road automation; unintended consequences of vehicle automation on driver’s workload, situation awareness and trust; and social interactions between driver, vehicle, and other road users. The following articles go further into the specific topics of interaction model/framework, control takeover, interaction with in-vehicle systems, impacts on gaze behaviors, and interface for visually impaired users. Janssen, Boyle, Kun, Ju, and Chuang (A Hidden Markov Framework to Capture Human-Machine Interaction in Automated Vehicles) introduce a probabilistic model/framework to formalize the beliefs that humans may have about the mode where a semi-automated vehicle is operating. They report three case studies. The framework provides a systematic and formal specification of mode confusion that can arise when the user fails to understand the mode of automation that is in operation at any given time. Favaro, Eurich, and Rizvi (“Human” Problems in Semi-AV: Understanding Drivers’ Reactions to Off-Nominal Scenarios) looked at the currently available data from AV field testing carried out in California from 2014 to 2017. The examination includes both qualitative and quantitative analyses. The findings include expected values for the response time, discussion of factors that affect dispersion, presentation of how to determine trust and experience effects in the data, as well as a comparison with state-of-the-art literature on the topic. In Kidd and Reagan (System attributes that influence reported improvement in drivers’ experiences with adaptive cruise control (ACC) and active lane keeping (ALK) after daily use in five production vehicles), 51 Insurance Institute for Highway Safety employees used an Audi A4 or Q7, Honda Civic, Infiniti QX60, or Toyota Prius for up to several weeks and completed surveys about their experiences. Drivers agreed that ACC improved driving experience more than ALK, but not true for all vehicles. Drivers were most comfortable using automation on interstates, and least comfortable using ACC on local roads and ALK on curvy roads. The study also indicated that automation should make smooth and gentle changes to steering or speed to improve perceptions of the technology and encourage use. Cramer and Klohr (Announcing Automated Lane Changes: Active Vehicle Roll Motions as Feedback for the Driver) studied active vehicle roll motions for announcing lane changes. Several designs were implemented and evaluated. 39 participants rated the vehicle roll motions for example regarding roll direction, intensity, usefulness, and the predictability of driving behavior. Results showed that active roll motions as feedback for announcing automated lane changes should be clearly perceptible and are considered useful, not misleading, and support the drivers regarding their mode/system awareness. Ahmadi and Machiani (Drivers’ Performance Examination Using a Personalized Adaptive Curve Speed Warning: Driving Simulator Study) developed an Adaptive Curve Speed Warning (ACSW) system and presented drivers a two-level visual and audio warning considering the variation in individual drivers’ perception-reaction time (PRT). The warning timing was adjusted according to a reward/punishment function. Drivers’ performance with ACSW and CSW (no PRT variation) were