Look Who’s Talking Now: Implications of AV’s Explanations on Driver’s Trust, AV Preference, Anxiety and Mental Workload

Abstract

This study proposes an autonomous vehicle incentive program design (AV-IPD) problem for a local government, which aims to promote the adoption of autonomous vehicles (AVs) by deploying AV lanes and subsidizing the purchase of AVs with the objective of system optimum subject to a fixed budget. As it is difficult to anticipate the changing AV market conditions in the future, we take into account the uncertainty in AV purchase price. For the AV-IPD problem, we firstly decide which regular lanes should be converted into exclusive AV lanes. Secondly, we determine the AV purchase subsidy for each realization of AV purchase price with a given AV lane deployment scheme. A binary logit model is applied to characterize the vehicle choice behavior of users. The AV-IPD problem is formulated as a two-stage stochastic programming model with equilibrium constraints. To solve the AV-IPD problem, we develop a solution method based on linear approximation techniques and duality theories. Numerical experiments are conducted to demonstrate the effectiveness of the proposed model and solution method. 1. Introduction The last few years have witnessed an explosive development on autonomous vehicle (AV) technology. Compared to the conventional vehicles (CVs), AVs can achieve a wide range of benefits in terms of traffic capacity, safety and vehicular emission. A successful implementation of AVs depends on the deployment of transport infrastructure catered for AVs and sufficient AV adoption. To promote the market penetration of AVs, a local government could design an AV incentive program to deploy exclusive AV lanes and subsidize users to purchase AVs subject to a fixed budget. The purpose of setting AV lanes is to separate AVs and CVs and exploit AV's advantage in improving road capacity. Current studies demonstrated that the road capacity could become up to triple in full AV environment by a shorter headway (Tientrakool et al., 2011). However, AVs become less beneficial under the mixed traffic with AVs and CVs. With a low AV ratio, the road capacity is possible to decrease due to speed variation and shock waves (Van Arem et al., 2006). To eliminate the issues arising from the mixed traffic, a local government can divide roads into regular lanes and AV lanes (Talebpour et al., 2017) so that AVs can travel with very short headway on AV lanes. The AV users have right of way on both regular and exclusive lanes while CV users can only travel on regular lanes. Different from conventional road construction projects, deploying AV lanes does not increase the number of lanes but converts some regular lanes into AV exclusive lanes. Some sensors are installed to facilitate the vehicle-to-vehicle (V2V) and vehicleto-infrastructure (V2I) communications (Jia et al., 2016). Some roads may require a novel traffic signal system (Li and Zhou, 2017). Besides, the users' vehicle choice between AVs and CVs is another feature needed to consider for the deployment of AV lanes. This is

Investigating explanations in conditional and highly automated driving: The effects of situation awareness and modality

“…For example, Körber et al (2018) provided explanations of the causes of takeover requests after the takeover transitions, which led to no decrease in trust or acceptance, but improved participants' understanding of system behaviors. Koo et al (2015) argued that explanations should be provided ahead of an event which also was supported by Haspiel et al (2018) andDu et. al.…”

Having the Right Attitude: How Attitude Impacts Trust Repair in Human—Robot Interaction

“…These efforts can rely on one of several strategies: apologies, denials, explanations, or promises [5], [11], [21]- [24]. Apologies are a communication and/or expression of remorse or regret [5], [8], [25]- [27]; denials are a rejection of culpability and often a redirection of blame [5], [11], [21]- [24]; explanations are clear and direct reasoning behind why a violation of trust occurred [1], [5], [8], [28]; and promises are statements conveying the intention to correct future behavior [8], [16], [22].…”

“I’d like an Explanation for That!”Exploring Reactions to Unexpected Autonomous Driving

“…In the automotive domain, too, explanations are used as a mean to foster trust in autonomous vehicles [11,33]. However, explaining autonomous driving systems to users comes with distinct challenges related to the driving infrastructure: In autonomous driving there are sources of information which may be unknown or hidden to the user, but which provide information explaining a decision by the autonomous vehicle.…”