Exploring Whether Managed Lanes Contribute to the Wide Adoption of the Connected and Autonomous Vehicles

Abstract

The rapid development of connected and autonomous vehicle (CAV) technology and the outperformance of the managed lane have drawn increasing attention with respect to wonder whether the CAV operating in the managed lane is helpful. Therefore, this study focuses on the methodology applied to explore whether the managed lane is beneficial in the gradual application of CAVs. Three managed lanes are considered, namely a bus lane, CAV dedicated lane, and high-occupancy vehicle lane. Given the unfamiliarity with the travel mode related to CAVs, the attitudinal variables for CAVs and their influencing relationship are designed and quantified by a structural equation model based on the extended technology acceptance model. A two-level nested logit model with different types of managed lanes as explanatory variables is further developed to analyze the travel choice of traditional modes and CAV modes. The difference in probabilities of travel choices under managed lanes is also explored. The results indicate that the three managed lanes have a positive influence on promoting the adoption of the CAV modes, with an increase of at least 13.5% shifted from traditional modes in the conventional lane. Moreover, the influences of managed lanes differ between transit CAVs and non-transit CAVs. The probabilities of choosing transit CAVs under three managed lanes increase in the range of 12.9%–27.2%, while those of non-transit CAVs only show a maximum 6.5% increase in the CAV dedicated lane. The study is expected to assist in designing differentiated development of CAVs with the deployment of managed lanes.

Keywords

high-occupancy vehicle lanes managed lanes mode choices operations performance measures planning and analysis

The connected and autonomous vehicle (CAV) represents a coming transformative mobility, which will arrive much earlier and more quickly than the majority who are accustomed to human-driven vehicles would think. In the current market, different degrees of automation have already been incorporated into existing cars, such as parking assist, adaptive cruise control, and collision avoidance systems. Studies also have revealed that this new technology has the potential to decrease traffic accidents ( 1 ), improve the stability of traffic streams ( 2 ), and reduce emissions ( 3 , 4 ).

To leverage this advantage fully and promote the application of CAVs, the influencing factors of whether to adopt CAVs or not has received much attention in recent studies ( 5 – 8 ), in which the CAV benefits (or perceived usefulness), CAV barriers (or perceived concerns), CAV perceived ease of use, pro-CAV attitude, and innovativeness ( 9 – 12 ) play important roles in choosing the travel mode related to CAV technology. However, because of the immaturity of CAV technology, barriers to CAV implementation and mass-market penetration remain. Although the advantages of CAVs over human-driven vehicles are prominent, the perceived risk, the perception of the difficulty of use, and the negative attitude about CAVs for the public are still difficult to thoroughly eliminate, which indicates that the CAV and the human-driven vehicle will remain in a mixed stage on the road for an extended period. Besides, there are studies that indicate that the potential efficiency of CAVs is significantly reduced when CAVs have a low penetration rate in the mixed traffic flow integrated with human-driven vehicles ( 13 ). Maximizing the advantages of CAVs under the technical limitations and the interference caused by the mixed traffic flow still needs attention.

Managed lanes, described by the U.S. Federal Highway Administration as highway facilities or a set of lanes where operational strategies are proactively implemented and managed in response to changing conditions, is vital for reducing congestion and offering drivers more reliable timeframes for getting to their destinations. The common types of managed lanes include bus lanes, high-occupancy vehicle (HOV) lanes, and other dedicated lanes. There are studies showing that the managed lane is capable of enhancing travel efficiency ( 14 – 16 ). Moreover, studies further conclude that CAVs operated in the managed lane have great potential to improve the traffic condition compared with roads without a CAV dedicated lane ( 17 ).

Considering these operational limitations of CAVs and potential advantages of managed lanes, implementing CAVs in the managed lane with respect to technology and management can be regarded as one of the promising effective solutions to maximize the socioeconomic benefits of CAVs. In this way, CAVs can form an efficient and energy-saving autonomous fleet, which can significantly highlight the advantages of connection and automation ( 18 , 19 ). The CAV dedicated lane, one type of managed lane, can also promote travel preferences for CAVs ( 20 ). However, the existing studies associated with CAVs deployed on a managed lane have focused on route choice and improvement of the throughput efficiency ( 21 ). The methodology applied to investigate the impact of the managed lane on the travel mode choice related to CAVs still remains to be explored.

In this context, this study focuses on the methodology application under the environment of CAVs implemented on managed lanes. Three types of managed lanes are considered, which are the bus lane, the CAV dedicated lane, and the HOV lane. Considering CAVs as an emerging technology, attitudinal variables related to CAVs considering their perception are investigated and quantified by a structural equation model (SEM) based on the extended technology acceptance model (TAM). The travel choice model integrated with CAV modes after the deployment of the managed lane is further developed. After that, the different impacts of three types of managed lanes are analyzed to obtain the targeted promotion strategy for adopting CAV modes.

The remainder of the paper is organized as follows. The second section offers the survey design and modeling framework, including the quantification of attitudinal variables and the estimated model of travel choice considering CAVs under scenarios with the deployment of managed lanes. The third section presents the survey design and discusses the estimated results and sensitivity analysis. The final section summarizes the main findings and recommendations.

Methodology

Survey Design

Given that CAV technology is evolving and has a low share in the current urban transport market, it is difficult to collect revealed preference (RP) data on mode choice. Comparatively, a stated preference (SP) survey could investigate travelers’ mode choices in hypothetical travel scenarios. Therefore, a SP survey has been employed in this study to collect data on travel choice in the context of CAVs deployed in different managed lanes.

Given that CAVs may still be unfamiliar to the public, this concept is described briefly before the questions. Images of non-transit CAVs and transit CAVs are shown to help respondents have an intuitive sense and understanding.

The questions in this survey consist of four parts.

The first part collected the history trip properties.

In the second part, respondents are asked for their mode choice under different trip scenarios combined with the deployment of managed lanes. Please note that there are two assumptions for this part. One is that the respondents are asked to imagine that they are in a hypothetical situation where all kinds of CAV modes and traditional travel modes are available in daily life, such as transit CAVs, private CAVs, taxi CAVs, and human-driven vehicles. Another assumption is that the respondents are required to make choices based on their most frequent travel purpose at present. Given that a conventional lane and three managed lanes are considered in this study, four scenarios related to different lanes are designed, and each scenario differs with respect to the lane type. In scenario 1, a conventional lane is provided, and all modes are permitted. In scenario 2, a CAV dedicated lane is provided that is only available for transit CAVs, private CAVs, and taxi CAVs. In scenario 3, a bus lane is provided, which is only open for the human-driven bus and transit CAVs. In scenario 4, a HOV lane is offered, which is reserved for the exclusive use of vehicles with two or more passengers. Briefly, five choices are presented in each scenario, including insisting on the current mode (which depends on their current transportation choice), transit CAV, private CAV, taxi CAV, and another mode. Take the scenario of the CAV dedicated lane as an example, the questionnaire interface for travelers who used to travel by private car is shown in Figure 1.

The third part is attitudinal questions about the individuals’ innovativeness, their attitude to CAVs, perceived benefits of CAVs, perceived concerns about CAVs, knowledge about CAVs, and their behavioral intention for CAVs.

For the last part, individual socioeconomic characteristics are collected.

Recently, the development of CAV technology has shown significant potential in China. There are nearly 700 well-known companies dedicated to the research and development of CAVs in China, such as Yutong and Baidu. The L4 autonomous shuttle bus, developed by Yutong, has already operated on open roads in Zhengzhou, Henan province, China. Moreover, according to a study published by McKinsey & Company, China will likely emerge as the world’s largest market for autonomous vehicles ( 22 ).

Figure 1.

Questionnaire interface for travelers used to traveling by the private car under the scenario of a connected and autonomous vehicle (CAV) dedicated lane.

Given those considerations, this survey was first conducted in December 2020 using an internet-based questionnaire among peer networks in China. During the two-week survey, 357 questionnaires were collected. After removing the questionnaires with an answer time of less than 120 s and the same answers to attitudinal questions, 345 valid samples were selected. The survey was conducted again in January 2023 to supplement the sample size further. During a more than two-week survey, 212 questionnaires were supplemented, of which 205 samples were valid. A total of 569 questionnaires were collected, and 550 samples remained for further modeling. Given that each questionnaire contains four scenarios, samples collected in this paper can be regarded as a narrow dataset. Finally, a total of 1936 survey data was used after being transformed from the narrow dataset to the wide dataset.

Model Specification

An overview of the methodology to investigate the impact of managed lanes on travel choice related to the CAV technology is presented. To capture the influence of attitudinal variables on choice behavior, partial attitudinal variables are considered as latent variables quantified by the SEM based on the extended TAM. After that, latent variables, scenario variables, socioeconomic variables, and travel variables are further added into the discrete choice model (DCM) determined by the nested logit (NL) model, which shows a significant improvement in model performance ( 23 ). Therefore, travel mode choices with the application of CAV technology in different managed lanes are analyzed to support appropriate investments for promoting large-scale CAV adoption. Figure 2 demonstrates the framework proposed in this study.

Figure 2.

Modeling framework.

Quantification of Attitudinal Variables: Structural Equation Model

The SEM is a modeling technique that can handle many endogenous and exogenous variables, as well as attitudinal variables (including latent variables) specified as linear combinations of the observed variables ( 24 ). The SEM is composed of a measurement model and a structural model. The measurement model estimates the relationship between attitudinal variables (including latent variables) and measurement indicators, which is estimated with a linear method. Note that since the attitude to CAVs and knowledge about CAVs in the attitudinal variables are reflected by only one measurement indicator, the latent variables included only the innovativeness, perceived benefits of CAVs, perceived concerns about CAVs, and behavioral intention for CAVs. The equation is shown in Equation 1:

y_{q} = Λ η_{p} + δ_{q}

(1)

where $y_{q} = q \times 1$ is a vector of attitudinal variables (including latent variables), $η_{p} = p \times 1$ is a vector of attitudinal variables (including latent variables), $Λ = q \times p$ is a coefficient matrix of factor loading between different attitudinal variables (including latent variables), and $δ_{q} = q \times 1$ is a vector of error terms subjecting to the independent and identically distributed multivariate normal.

The structural model explores the structural relationship within the latent variables and between exogenous variables and latent variables (including the innovativeness, perceived benefits of CAVs, perceived concerns about CAVs, and behavioral intention for CAVs), which is estimated with a linear method. The equation is expressed in Equation 2:

η_{p} = Γ x_{q} + ς_{p}

(2)

where $x_{q} = q \times 1$ is a vector of observed exogenous variables, including age, income, education, and so forth, $Γ = p \times q$ is a coefficient matrix of the regression effects of $x_{q}$ on the latent variables, and $ς_{p} = p \times 1$ is a vector of error terms subjecting to the independent and identically distributed multivariate normal.

The TAM was originally developed to analyze the acceptance of information technology innovations but has since been applied to the transportation field, including for quantitative analysis of the acceptance of electric vehicles ( 25 ) and environmentally friendly transportation systems ( 26 ). In light of CAVs being a new emerging technology, an extended TAM is further developed to capture influencing relationships among attitudinal variables. As attitudes toward use influenced by the perceived usefulness and the perceived ease of use are recognized as an influencing factor of the intention to use in the TAM model ( 27 ), the perceived usefulness is expressed as the perceived benefits of CAVs with a relationship with the perceived concerns about CAVs, while the perceived ease of use is represented by two attitudinal variables, namely the innovativeness and the knowledge about CAVs in the extended TAM model. In addition, these variables are further assumed to have an effect on the behavior intention for CAVs. Simultaneously, to find out how individual attributes affect attitudinal variables, individual differences are also taken into account. Figure 3 displays the SEM framework based on the extended TAM. Here, Hi (© =1, 2, 3, …, 13) represents a hypothetical relationship. Take H1 as an example, which means that the innovativeness has a significant effect on the behavioral intention for CAV.

Figure 3.

Structural equation modeling framework based on the extended technology acceptance model (TAM).

Estimation of Travel Choice: Nested Logit Model

A two-level NL model is proposed to investigate the impact of managed lanes on travel choice related to CAV technology in this study. With respect to whether travelers choose the travel mode related to CAV technology because of the deployment of managed lanes, the upper level is set as two nests, a traditional mode nest and a CAV mode nest, respectively. Moreover, considering that travelers’ trust in the transit CAV may differ from that of the non-transit CAV, the lower nest is further divided into transit and non-transit modes for the traditional mode nest and transit CAV and non-transit CAV in the CAV mode nest. It is noteworthy that five traditional modes (transit, private car, taxi, bike, and walk) and three modes related to CAV technology (private CAV, transit CAV, and CAV taxi) are included in the survey designed in this paper. With respect to the sample size limitation and many estimated parameters, travel modes other than transit are uniformly classified as the non-transit mode. Specifically, scenario attributes from the perspective of the deployment of four lanes are considered as explanatory variables. Finally, the nest structure in the NL model is shown in Figure 4.

Figure 4.

Nest structure in the nested logit model.

This NL structure includes four alternatives numbered as $i$ = 1, 2, 3, 4, which are categorized into two nests, $m$ =1, 2. According to the theory of random utility, the utility $U_{in}$ that a traveler $n$ chooses particular mode $i$ is shown in Equation 3:

U_{in} = V (x_{in}, η_{in}; β_{i}) + ε_{in}

(3)

where $V$ is the deterministic component of the utility function, $x_{in}$ is the observed variable, $η_{in}$ is the attitudinal variables measured by the SEM, $β_{i}$ is the coefficient to be estimated associated with the corresponding variable, and $ε_{in}$ is the stochastic term that is subject to Gumbel distribution.

The term $P_{n} (rm)$ represents the joint probability that traveler $n$ chooses lower level alternative $i$ under nests $m$ , is shown in Equation 4:

P_{n} (rm) = P_{n} (r | m) P_{n} (m), r \in R_{mn}

(4)

where $P_{n} (r | m)$ is the conditional probability that individual $n$ chooses $r$ in the nest $m$ , $P_{n} (m)$ is the probability that individual $n$ chooses $m$ , $R_{mn}$ is the set of mode choices for individual $n$ , and $P_{n} (r | m)$ is determined in the lower level multinomial logit (MNL) model as in Equation 5:

P_{n} (r | m) = \frac{\exp (μ_{m} V_{(r | m) n})}{\sum_{i = 1} \exp (μ_{m} V_{(r' | m) n})}

(5)

where $V_{(r | m) n}$ is the deterministic component of the utility function of alternative $r$ under $m$ and $μ_{m}$ is the scale parameter for the lower level, set as 1 for the purpose of simple estimation ( 28 ).

The natural logarithm, $V_{mn}^{*}$ , of the denominator is taken as the expected maximum utility in set $R_{mn}$ as in Equation 6:

V_{mn}^{*} = \frac{1}{μ_{m}} In \sum_{r' \in M_{n}} \exp (V_{(r' | m) n})

(6)

In the NL model, each choice is conditioned on the suppositional nest, which is influenced by the logsum ( 29 ). The logsum is computed for the nest and added to the upper level of the model as the explanatory value. Thus, the probability that traveler $n$ chooses nest m $P_{n} (m)$ can be expressed as in Equation 7:

P_{n} (m) = \frac{\exp (λ_{m} (V_{mn} + V_{mn}^{*}))}{\sum_{m' \in M_{n}} \exp (λ_{m} (V_{mn} + V_{m' n}^{*}))}

(7)

where $V_{mn}$ is the deterministic component of the utility function, which varies only with nest $m$ and traveler $n$ , $λ_{m}$ is the upper level scale parameter to be calibrated, indicating different distributions of the disturbances between different nests and levels, and $λ_{m}$ is the scale parameter, which should be 0–1, indicating that the assumption of the nest structure is rational ( 30 ).

Analysis of Results

Sample Analysis

Descriptive statistics of categorical attributes are presented in Table 1. The respondents are evenly distributed across age groups with a high level of education. More than 50% of the travel purpose is commuting. The proportions of survey participants who travel by transit and private cars are relatively balanced.

Table 1

Descriptive Statistics of Categorical Attributes (Sample Size = 550)

Attribute	Description	Frequency	Ratio
Gender	Male	273	49.64%
Gender	Female	277	50.36%
Age	Age 18–25	179	32.55%
	Age 26–35	124	22.55%
	Age 36–45	121	22.00%
	Age 46 or older	126	22.91%
Education	Below secondary technical school	132	24.00%
	Undergraduate	258	46.91%
	Postgraduate	128	23.27%
	PhD and above	32	5.82%
Monthly income	<¥5000	269	48.91%
	¥5000–10,000	174	31.64%
	>¥10,000	107	19.45%
Number of car owners	Null	243	44.18%
	1 vehicle	257	46.73%
	2 vehicles	45	8.18%
	3 and more vehicles	5	0.91%
Main travel purpose	Commuting	290	52.73%
	Leisure/entertainment/shopping	181	32.91%
	Others	79	14.36%
History travel mode under main travel purpose	Private car	221	40.18%
	Taxi or online car-hailing	39	7.09%
	Motorcycle or electric bike	41	7.45%
	Walking or bicycle	41	7.45%
	Transit (bus and metro)	208	37.82%
Travel time	<15 min	64	11.64%
	15–30 min	172	31.27%
	30–45 min	149	27.09%
	45–60 min	84	15.27%
	>60 min	81	14.73%
Travel frequency per week	<2 times	113	20.55%
	3–4 times	101	18.36%
	5–6 times	82	14.91%
	>6 times	254	46.18%
Driving year	<1 year	194	35.27%
	1–3 year	100	18.18%
	3–6 year	73	13.27%
	>6 year	183	33.27%

Considering that prior travel experience related to CAVs or managed lanes may influence travel choices, questions related to the advanced driver assistance system (ADAS), bus lane, and HOV lane are set. The survey results are shown in Figure 5. It can be seen that only 23% of respondents are familiar with the ADAS. More than 50% had experienced with the bus lane, but only 26% experienced the HOV lane. The results identified that CAV technology is still unfamiliar to the public, and the bus lane is more common than the HOV lane, which is consistent with the development status of the case study area.

Figure 5.

Descriptive statistics of the travel experience related to the connected and autonomous vehicle or managed lane.

Descriptive statistics of attitudinal attributes on a five-point Likert scale are presented in Table 2. As shown in Table 2, there are six attitudinal attributes (innovativeness, perceived benefits of CAVs, perceived concerns about CAVs, knowledge of CAVs, attitude to CAVs, and behavioral intention for CAVs) and 12 questions are asked. Ind01–Ind03 ask respondents about innovativeness. Ind04–Ind06 focus on their perceived benefits for CAVs. Ind07 and Ind08 are set to obtain respondents’ concerns about CAVs. Ind09 and Ind10 ask about their knowledge and development attitude toward CAVs, respectively. Ind11 and Ind12 aim to learn about respondents’ behavioral intention to use CAVs. Table 2 also includes the percentage of responses for all Likert scale levels of indicator variables. In summary, respondents have a positive attitude toward the development and application of CAVs in the future. Besides, more than 70% of the respondents have at least certain knowledge of CAVs to some extent, in which 32% of respondents are very familiar with CAV technology.

Table 2.

Descriptive Statistics of the Likert Attributes (Sample Size = 550)

Item	Likert scale variables	Likert scale levels: strongly disagree (1) to strongly agree (5)					Mean	SD
Item	Likert scale variables	1	2	3	4	5	Mean	SD
Ind01: Innov1	I trust and am willing to try new objects, such as new technology products.	1%	2%	20%	53%	24%	3.97	0.78
Ind02: Innov2	When a new technology product comes into the market, I would like to be the first to use it.	1%	6%	36%	41%	16%	3.66	0.84
Ind03: Innov3	I can quickly master how to use new products.	1%	4%	24%	50%	21%	3.87	0.80
Ind04: Benefit1	I believe that CAVs can effectively avoid the mistakes of human driving and reduce traffic accidents.	1%	3%	38%	46%	11%	3.63	0.76
Ind05: Benefit2	I believe that CAVs can effectively alleviate road traffic congestion.	1%	3%	33%	47%	16%	3.72	0.82
Ind06: Benefit3	I believe that CAV technology is safer than human driving.	2%	11%	48%	29%	10%	3.34	0.87
Ind07: Concern1	I am concerned about CAVs cannot guarantee my property and personal safety.	3%	17%	45%	30%	5%	3.17	0.86
Ind08: Concern2	I am concerned about collisions and crashes when CAVs run on the same lanes with regular vehicles.	4%	16%	41%	32%	7%	3.21	0.94
Ind09: CAV knowledge	I have a good knowledge of CAV technology, such as use cost, advantages over traditional vehicles on safety enhancement, congestion relief, comfort improvement, etc.	6%	21%	42%	26%	6%	3.03	0.97
Ind10: CAV attitude	I am positive toward the development and application of CAV technology in the future.	1%	3%	30%	52%	14%	3.77	0.75
Ind11: Intention1	I would like to be the first to experience CAVs.	1%	9%	36%	44%	10%	3.54	0.82
Ind12: Intention2	I would like to recommend friends around me to travel with CAVs.	1%	8%	44%	37%	9%	3.43	0.82

Note: SD = to standard deviation; CAV = connected and autonomous vehicle.

Impact Analysis of Attitudinal Variables Related to CAVs

Given that partial attitudinal variables could not be directly captured, four attitude variables are considered as latent variables (innovativeness, perceived benefits of CAVs, perceived concerns about CAVs, and behavioral intention for CAVs) and quantified firstly by the SEM based on the extended TAM. The confirmatory factor analysis of the proposed model is performed to analyze the validity and reliability of the quantification of latent variables using the software Mplus. The fit indices are shown in Tables 3 and 4.

Table 3.

Convergent Validity

Latent variable	Item	SFL (>0.5)	SE	CR (>0.7)	AVE (>0.5)
Innovativeness	Ind01: Innov1	0.78	0.01	0.83	0.62
	Ind02:Innov2	0.84	0.01
	Ind03: Innov3	0.74	0.01
Perceived benefits of CAVs	Ind04: Benefit1	0.84	0.01	0.82	0.61
	Ind05: Benefit2	0.73	0.01
	Ind06: Benefit3	0.77	0.01
Perceived concerns about CAVs	Ind07: Concern1	0.78	0.02	0.79	0.65
Perceived concerns about CAVs	Ind08: Concern2	0.83	0.02	0.79	0.65
Behavioral intention for CAVs	Ind11: Intention1	0.90	0.01	0.86	0.76
Behavioral intention for CAVs	Ind12: Intention2	0.84	0.01	0.86	0.76

Note: Goodness-of-fit indices: χ²/df = 2.17; CFI = 0.953; TLI = 0.914; RMSEA = 0.05; SRMR = 0.028. CAV = connected and autonomous vehicle; SFL = standardized factor loading; CR = construct reliability; AVE = average variance extracted; SE = standard error.

Table 4.

Correlations Among Latent Variables

Latent variables	Innovativeness	Perceived benefits of CAVs	Perceived concerns about CAVs	Behavioral intention for CAVs
Innovativeness	0.79	0.48	0.01	0.56
Perceived benefits of CAVs	0.48	0.78	−0.09	0.57
Perceived concerns about CAVs	0.01	−0.09	0.81	−0.21
Behavioral intention for CAVs	0.56	0.57	−0.21	0.87

Note: The values in the diagonal (set in bold) are the square roots of average variance extracted (AVE); others are the correlation among different latent variables. CAV = connected and autonomous vehicle.

As shown in Table 3, the convergent validity is supported by the standardized factor loadings (SFLs), construct reliability (CR), and average variance extracted (AVE). All the values of SFL meet the minimum criterion of 0.5. The values of CR range from 0.79 to 0.86, exceeding the critical value of 0.7, which indicates a satisfactory estimation in general ( 31 , 32 ). The AVE of all latent variables ranges between 0.61 and 0.76, all above the suggested value of 0.5 ( 31 , 32 ). Besides, the model fit is assessed with the chi-square/df (χ²/df) statistics, the comparative fit index (CFI), the Tucker–Lewis index (TLI), the root mean square of approximation (RMSEA), and the standardized root mean square residual (SRMR). The value of χ²/df between 1 and 3 and the values of CFI and TLI greater than 0.9 indicate acceptability, except for the RMSEA and SRMR, for which values lower than 0.08 indicate an acceptable fit to the data ( 31 , 32 ).

Moreover, discriminant validity distinguishes one latent variable from the others. As shown in Table 4, it is suggested that discriminant validity is achieved if the squared root of AVE for each latent variable is greater than the correlation coefficients of the corresponding inter-variables ( 33 ). The values in the diagonal are more significant than the value of corresponding inter-variables. All the indexes confirmed the validity and fitness of the proposed SEM.

The estimated results of the SEM based on the extended TAM considering the impact of attitudinal variables related to CAVs are conducted using the software Mplus. The hypothetical relation verification and impact quantification are presented in Figure 6.

Figure 6.

Estimated results of relationships.

As shown in Figure 6, 13 proposed hypothetical relations are supported significantly, which proves the rationality of the extended TAM proposed in this study. The results show that the innovativeness and the perceived benefits of CAVs have a positive effect on the behavioral intention for CAVs, while the perceived concerns about CAVs show a negative impact on the behavioral intention for CAVs, as found by Adnan et al. ( 34 ), Nastjuk et al. ( 35 ), and Zhu et al. ( 36 ). Besides directly improving the behavioral intention for CAVs, knowledge about CAVs can promote the willingness to use CAVs by reducing perceived concerns about CAVs. Furthermore, the innovativeness and knowledge about CAVs have a positive effect on the perceived benefits of CAVs, indicating that individuals with innovativeness and greater knowledge of CAVs are more likely to perceive the usefulness of CAV technology. Thus, it is essential to enhance the dissemination of CAV knowledge and its benefits to the public.

For the sake of brevity, only the significant impacts between socioeconomic attributes and attitudinal variables are presented in Table 5. The population with high education has more knowledge about CAVs, as in Haboucha et al. ( 37 ) and Huang et al. ( 38 ). Besides, individuals aged 26–35 have a significant positive impact on the willingness to learn and use CAVs, which corresponds to the results of Lavieri et al. ( 39 ). This phenomenon indicates that young adults aged 26–35 would be the early adopters of CAVs. Moreover, the samples with a master’s degree and above have more significant perceived concerns about CAVs, with values of 0.097 and 0.046, respectively, which are consistent with previous studies ( 40 , 41 ). The travel experience of ADASs, which is related to CAV technology, contributes to the perceived benefits of CAVs, knowledge about CAVs, and behavioral intention for CAVs.

Table 5.

Estimated Results of Travel Choice Related to Connected and Autonomous Vehicles (CAVs) under the Managed Lane

		Coefficient (p-value)^Significance
Explanatory variables		Innovativeness	Attitude to CAVs	Perceived benefits of CAVs	Perceived concerns about CAVs	Knowledge about CAVs	Behavioral intention for CAVs
Gender	Male	-	−0.050***	-	-	−0.064***	-
Age (46–55 as the reference)	18–25	0.160***	0.134***	-	−0.089**	0.181***	0.049*
	26–35	0.102**	0.148***	-	−0.130***	0.233***	0.188***
	36–45	0.094*	-	-	-	0.232***	0.140***
Education (secondary technical school as the reference)	Under-graduate	-	−0.199***	0.050*	-	0.112***	−0.062***
	Post-graduate	0.059*	−0.316***	0.036*	0.097***	0.106***	−0.056**
	PhD and above	0.065**	−0.207***	0.063***	0.046*	0.140***	−0.042**
Monthly income (<¥5000 as the reference)	¥5,000-¥10,000	0.160***	0.074***	0.053**	-	-	-
Monthly income (<¥5000 as the reference)	>¥10,000	0.047*	0.232***	0.129***	-	0.074***	-
Experience of ADAS		0.228***	0.275***	0.050**	−0.205***	0.205***	0.053**
Driving year		-	-	-	-	0.154***	−0.048*

Note: Significance levels: - not significant, *0.1, **0.05, ***0.01. ADAS = advanced driver assistance system.

Travel Choice Related to CAVs Considering the Impact of Managed Lanes

Considering the socioeconomic attributes, travel variables, attitudinal variables related to CAVs, and scenario variables concerning managed lanes, the proposed NL model is programmed and calibrated. The parameters are estimated based on the maximum likelihood method using Python Biogeme ( 42 ). Since transit is accessible for all travelers, the transit mode is used as the base (reference) choice in this study. The estimated results are enumerated in Table 6.

Table 6.

Estimated Results of Travel Choice Related to Connected and Autonomous Vehicles (CAVs) under the Managed Lane (N = 1936)

Utility function for mode	Explanatory variables		Coefficient
Transit	Alternative-specific constant		2.630***
Non-transit	Socioeconomic variables
	Age (46–55 as the reference)	18–25	1.450***
		26–35	2.020***
		36–45	1.920***
	Monthly income (<¥5000 as the reference)	¥5000–10,000	1.120***
	Monthly income (<¥5000 as the reference)	>¥10,000	1.570***
	Number of car owners (constant variable)		0.120*
Transit CAVs	Alternative-specific constant		1.350**
	Socioeconomic variables
	Age (46–55 as the reference)	18–25	0.407***
	Travel properties
	History travel mode (non-motorized mode as the reference)	Public transportation	−0.706***
	Main travel purpose (commute as the reference)	Leisure/entertainment/shopping	0.429***
	Latent variables
	Perceived benefits of CAVs		0.106**
	Behavioral intention for CAVs		0.214***
	Scenario attributes (conventional lane as the reference)
	Deployment of bus lane		1.310***
	Deployment of CAV dedicated lane		1.100***
	Deployment of HOV lane		0.722***
Non-transit CAVs	Alternative-specific constant		2.110*
	Socioeconomic variables
	Age (46–55 as the reference)	18–25	0.510**
	Age (46–55 as the reference)	26–35	0.417*
	Monthly Income (<¥5000 as the reference)	¥5000–10,000	0.433***
	Monthly Income (<¥5000 as the reference)	> ¥10,000	0.541***
	Driving year (constant variable)		0.013*
	Travel properties
	Travel frequency peer week (<3 as the reference)	5–6	0.433**
	Travel frequency peer week (<3 as the reference)	>6	0.277*
	History travel mode (non-motorized mode as the reference)	Private car	1.580***
	History travel mode (non-motorized mode as the reference)	Taxi or online car-hailing	0.975***
	Latent variables
	Innovativeness		0.083*
	Attitude to CAVs		0.298***
	Perceived concerns about CAVs		−0.074*
	Knowledge about CAVs		0.191***
	Scenario attributes (conventional lane as the reference)
	Deployment of bus lane		−0.021*
	Deployment of CAV dedicated lane		0.908***
	Deployment of HOV lane		0.388**
Scale parameters			0.770
Adjusted McFadden’s rho-square			0.220
Final log-likelihood			−2088

Note: significance levels: *0.1, **0.05, ***0.01. HOV = high-occupancy vehicle.

As shown in Table 6, the proposed NL model fits well in the analysis of travel choice related to CAVs considering the impact of the managed lane, with the scale parameters between 0 and 1, the adjusted McFadden’s rho-square between 0.2 and 0.4, and the relatively high final log-likelihood.

Individuals aged between 18 and 35, with an income above ¥5000, traveling more than five times per week, and used to travel by non-transit modes (private car and taxi or online car-hailing), are more inclined to choose non-transit CAVs, which indicates that lower age, higher income, and frequency of travel all show a positive impact on the adoption of non-transit CAVs. The effects of income and travel frequency correspond with the literature ( 43 , 44 ). However, according to the studies of Abraham et al. ( 45 ) and Hassan et al. ( 46 ), older adults are more concerned about non-transit CAVs than younger adults in Canada. One of the possible major reasons for the difference in the age effect is the land use characteristic around the living place. As for the choice of transit CAV, it is interesting to find that the travel habit of using public transportation shows a significant influence, which may be because the travel indolence is caused by the psychology of loss aversion for new technology. Compared with commuting, individuals with the travel purposes of leisure, entertainment, and shopping tend to be more likely to try transit CAVs. Wadud and Mattioli ( 47 ) and Yueshuai He et al. ( 48 ) also suggested that work trips contributed to a larger probability of finding ownership of a household CAV. This result indicates that transit routes with CAV technology allocated between tourist parks, shopping malls, and other recreational places may achieve more considerable benefits than expected.

The impacts of attitudinal variables related to CAVs are also obtained. As expected, innovativeness, positive attitude to CAVs, and behavioral intention for CAVs have a positive influence on the travel choice of the CAV mode, while perceived concerns about CAVs contribute to eliminating the tendency to use CAVs, consistent with the findings of Zhu et al. ( 36 ), Benleulmi and Ramdani ( 49 ), and Zhang et al. ( 50 ). Moreover, the negative parameter of the perceived concerns about CAVs is presumably because of misunderstanding caused by frequently eye-catching information about accidents involving immature CAVs.

Different types of managed lanes are introduced into the NL model to investigate these impacts on the travel choice related to CAVs. Although most managed lanes show a positive influence on promoting the adoption of the travel mode related to CAVs, the influences still differ between different travel modes. As for non-transit CAVs, the promotion effect of the CAV dedicated lane is more noticeable than the HOV lane. In contrast, the deployment of the bus lane affects most of the travel choices of transit CAVs, with the CAV dedicated lane following closely behind.

To investigate the impact of different types of managed lanes explicitly, the choice probabilities of travel modes, including that related to CAV technology, in three managed lanes are further estimated. The travel choice in the conventional lane is also obtained for comparison. The estimated results are shown in Figure 7.

Figure 7.

Choice probabilities of travel modes: (a) probability for the conventional lane, (b) probabilities for the bus lane, (c) probability for the connected and autonomous vehicle (CAV) dedicated lane, and (d) probability for the high-occupancy vehicle (HOV) lane.

As shown in Figure 7, all three managed lanes mentioned in this paper promote travel modes associated with CAV technology compared to the conventional lane. It can be seen that the total choice probabilities of the CAV mode for the conventional lane, bus lane, CAV dedicated lane, and HOV lane are 59.3%, 76.2%, 80.0%, and 72.8%, respectively. Compared to the probability of choosing the CAV mode for the conventional lane, the corresponding probabilities for the bus lane, CAV dedicated lane, and HOV lane are increased by 16.9%, 20.7%, and 13.5%, respectively, which means that at least 13.5% of travelers who choose the traditional mode for the conventional lane prefer to use the CAV mode if the managed lanes are implemented. Moreover, the promotion effect of the CAV mode is most pronounced in the CAV dedicated lane, followed by the bus lane, and last is the HOV lane.

The promotion effect of different managed lanes differs for transit CAVs versus non-transit CAVs. For transit CAVs, the bus lane among the different managed lanes is still one of the infrastructures to most likely improve the choice probability of public transit, with 27.0% in the conventional lane, 54.2% in the bus lane, 41.2% in the CAV dedicated lane, and 39.9% in the HOV lane. The probability of travel choice for transit CAVs increases by 12.9%–27.2% after the implementation of managed lanes. However, the choice probabilities for non-transit CAVs in the conventional lane, bus lane, CAV dedicated lane, and HOV lane are 32.3%, 22.0%, 38.8%, and 32.9%, respectively. The bus lane shows a significant adverse effect on non-transit CAVs, while the CAV dedicated lane and HOV lane boost the choice probability of non-transit CAVs by 6.5% (the difference between 38.8% and 32.3%) and 0.6% (the difference between 32.9% and 32.3%), respectively. This phenomenon exists probably because the bus lane overcrowds the limited road resources for non-transit CAVs. Given the above analysis, the integrated development of CAV technology and the bus lane shows significant potential to improve traffic congestion and enhance the attractiveness of public transportation, especially for areas with more constrained traffic resources.

Sensitivity Analysis

Travelers’ attitudes toward travel modes related to CAV technology are critical to travel choices. Based on the six attitudinal variables that have been quantified, the impact of changes in these variables on travel choice is further explored. The changing trends are shown in Figure 8.

Figure 8.

Sensitivity analysis of attitudinal variables: (a) transit, (b) non-transit, (c) transit connected and autonomous vehicle (CAV), and (d) non-transit CAV.

As shown in Figure 8, the improvement of all attitude variables, except the perceived concerns about CAVs, has a promoting effect on the mode choice related to CAV technology. The attitude to CAVs shows the most significant effect on boosting transit CAVs, while knowledge about CAVs has the highest positive effect on non-transit CAVs. An interesting phenomenon is that perceived concerns about CAVs are a more sensitive factor in the travel choice compared to the perceived benefits of CAVs, which suggests that it would be more helpful to eliminate social concerns about CAVs than to vigorously promote the advantages of CAVs above human-driven vehicles in gradually promoting the CAV to the market.

Conclusions and Recommendations

Given the relatively closed operating conditions, the managed lane shows the potential to provide a broader application scenario for the promotion of CAVs, which is limited to immature technology and public concerns about operating smoothly in mixed traffic flow. This study focuses on the methodology application under the provision of CAVs potentially implemented on managed lanes. Three managed lanes are considered, namely a bus lane, CAV dedicated lane, and HOV lane. In addition, six attitudinal variables related to CAV perceptions and the internal relationships between them are quantified. After that, a two-level NL model incorporating managed lanes as explanatory variables is proposed. The choice probabilities of travel modes related to CAV technology under different managed lanes are further explored to analyze the changes in travel choice with the impact of the deployment of managed lanes. The main conclusions are as follows.

The attitudinal variables of the innovativeness, positive attitude to CAVs, and behavioral intention for CAVs show a positive influence on mode choice related to CAV technology (transit CAVs and non-transit CAVs). Moreover, the perceived concerns about CAVs are more sensitive in the mode choice compared to the perceived benefits of CAVs, which suggests that eliminating social concerns about CAVs is more helpful than strongly advocating the outperformance of CAVs above human-driven vehicles during the promotion of CAVs to the market.

At least 13.5% of travelers choosing traditional modes in the conventional lane prefer the mode related to the CAV technology after the managed lanes are deployed. Besides, the most effective in promoting travelers to shift from traditional modes to CAV modes are the CAV dedicated lane, followed by the bus lane, and last is the HOV lane.

The impact of different managed lanes differs between transit CAVs and non-transit CAVs. The probabilities of choosing transit CAVs under three managed lanes increase in the range of 12.9%–27.2% compared to that of the conventional lane. In contrast, the CAV dedicated lane and HOV lane boost the choice probability of non-transit CAVs by 6.5% and 0.6%, respectively, while the bus lane shows a negative impact, presumably because of crowded road resources.

This study provides an insight for understanding the impact of a managed lane on travel choice with the mode related to CAV technology. It is acknowledged that SP survey results may not completely reflect the situation when the CAV reaches a high penetration rate; however, the proposed model can be further modified according to the actual situation. In the future, a larger sample size and heterogeneous models are also recommended to refine the CAV travel choice mechanism.

Footnotes

Acknowledgements

Thanks are given to the anonymous reviewers for their helpful comments on this paper.

Author Contributions

The authors confirm their contribution to the paper as follows: study conception and design: J. Ma, Z. Tian, X. Chen, L. Yu; data collection: Z. Tian; analysis and interpretation of results: J. Ma, Z. Tian, X. Chen, L. Yu; draft manuscript preparation: J. Ma, Z. Tian, X. Chen. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: this paper was supported by the National Natural Science Foundation of China (NSFC) under grants #71871013 and #72271020.

ORCID iDs

Xumei Chen

Lei Yu

Data Accessibility Statement

The datasets generated during and/or analyzed during the current study are not publicly available because of the limitation of personal privacy but are available from the corresponding author on reasonable request.

References

Fagnant

D. J.

Kockelman

Preparing a Nation for Autonomous Vehicles: Opportunities, Barriers and Policy Recommendations. Transportation Research Part A: Policy and Practice, Vol. 77, 2015, pp. 167–181.

Talebpour

Mahmassani

H. S.

Influence of Connected and Autonomous Vehicles on Traffic Flow Stability and Throughput. Transportation Research Part C: Emerging Technologies, Vol. 71, 2016, pp. 143–163.

Wadud

MacKenzie

Leiby

Help or Hindrance? The Travel, Energy and Carbon Impacts of Highly Automated Vehicles. Transportation Research Part A: Policy and Practice, Vol. 86, 2016, pp. 1–18.

Eilbert

Noel

O’Donnell

Smith

Evaluating Energy and Emissions Impacts of Cooperative Adaptive Cruise Control (CACC) Technology Through Traffic Microsimulations. Proc., Automated Vehicles Symposium, SanFrancisco, 2017.

Bansal

Kumar

R. R.

Raj

Dubey

Graham

D. J.

Willingness to Pay and Attitudinal Preferences of Indian Consumers for Electric Vehicles. Energy Economics, Vol. 100, 2021, p. 105340.

Wang

Salehin

M. F.

Habib

K. N.

A Discrete Choice Experiment on Consumer’s Willingness-to-Pay for Vehicle Automation in the Greater Toronto Area. Transportation Research Part A: Policy and Practice, Vol. 149, 2021, pp. 12–30.

Nair

G. S.

Astroza

Bhat

C. R.

Khoeini

Pendyala

R. M.

An Application of a Rank Ordered Probit Modeling Approach to Understanding Level of Interest in Autonomous Vehicles. Transportation, Vol. 45, 2018, pp. 1623–1637.

Haboucha

C. J.

Ishaq

Shiftan

User Preferences Regarding Autonomous Vehicles. Transportation Research Part C: Emerging Technologies, Vol. 78, 2017, pp. 37–49.

Manfreda

Ljubi

Groznik

Autonomous Vehicles in the Smart City Era: An Empirical Study of Adoption Factors Important for Millennials. International Journal of Information Management, Vol. 58, 2019, p. 102050.

10.

Lee

Mulrow

Haboucha

C. J.

Derrible

Shiftan

Attitudes on Autonomous Vehicle Adoption Using Interpretable Gradient Boosting Machine. Transportation Research Record: Journal of the Transportation Research Board, 2019. 2673: 865–878.

11.

Leicht

Chtourou

Youssef

K. B.

Consumer Innovativeness and Intentioned Autonomous Car Adoption. The Journal of High Technology Management Research, Vol. 29, No. 1, 2018, pp. 1–11.

12.

Waung

McAuslan

Lakshmanan

Trust and Intention to Use Autonomous Vehicles: Manufacturer Focus and Passenger Control. Transportation Research Part F: Traffic Psychology and Behaviour, Vol. 80, 2021, pp. 328–340.

13.

Chen

Ahn

Chitturi

Noyce

D. A.

Towards Vehicle Automation: Roadway Capacity Formulation for Traffic Mixed with Regular and Automated Vehicles. Transportation Research Part B: Methodological, Vol. 100, 2017, pp. 196–221.

14.

Tan

Gao

H. O.

Hybrid Model Predictive Control Based Dynamic Pricing of Managed Lanes with Multiple Accesses. Transportation Research Part B: Methodological, 2018, Vol. 112, pp. 113–131.

15.

Burris

M. W.

Brady

J. F.

Unrevealed Preferences: Unexpected Traveler Response to Pricing on Managed Lanes. Transportation Research Record: Journal of the Transportation Research Board, 2018. 2672: 23–32.

16.

Guensler

Kim

Khoeini

Factors Affecting Atlanta Commuters’ High Occupancy Toll Lane and Carpool Choices. International Journal of Sustainable Transportation, Vol. 14, No. 12, 2020, pp. 932–943.

17.

Wang

Influence of Exclusive Lanes for Connected and Autonomous Vehicles on Freeway Traffic Flow. IEEE Access, Vol. 7, 2019, pp. 50168–50178.

18.

Zhong

Lee

The Effectiveness of Managed Lane Strategies for the Near-Term Deployment of Cooperative Adaptive Cruise Control. Transportation Research Part A: Policy and Practice, Vol. 129, No. 9, 2019, pp. 257–270.

19.

Guo

Leveraging Existing High-Occupancy Vehicle Lanes for Mixed-Autonomy Traffic Management with Emerging Connected Automated Vehicle Applications. Transportmetrica A: Transport Science, Vol. 16, No. 3, 2020, pp. 1375–1399.

20.

Shabanpour

Golshani

Shamshiripour

Mohammadian

A. (Kouros)

. Eliciting Preferences for Adoption of Fully Automated Vehicles Using Best-Worst Analysis. Transportation Research Part C: Emerging Technologies, Vol. 93, No. 7, 2018, pp. 463–478.

21.

Yamamoto

Impact of Dedicated Lanes for Connected and Autonomous Vehicle on Traffic Flow Throughput. Physica A: Statistical Mechanics and its Applications, Vol. 512, 2018, pp. 588–597.

22.

McKinsey Center for Future Mobility. China May Become the World’s Largest Autonomous Vehicle Market. McKinsey&Company. 2018. https://www.mckinsey.com.cn/.

23.

Walker

Ben-Akiva

Generalized Random Utility Model. Mathematical Social Sciences, Vol. 43, No. 3, 2002, pp. 303–343.

24.

Golob

T. F.

Structural Equation Modeling for Travel Behavior Research. Transportation Research Part B: Methodological, Vol. 37, No. 1, 2003, pp. 1–25.

25.

Dudenhöffer

Why Electric Vehicles Failed: An Experimental Study with PLS Approach Based on the Technology Acceptance Model. Journal of Management Control, Vol. 24, 2013, pp. 95–124.

26.

Chen

Green Helpfulness or Fun? Influences of Green Perceived Value on the Green Loyalty of Users and Non-Users of Public Bikes. Transport Policy, Vol. 47, 2016, pp. 149–159.

27.

Davis

F. D.

Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly, Vol. 13, No. 3, 1989, pp. 319–340.

28.

Hunt

G. L.

Alternative Nested Logit Model Structures and the Special Case of Partial Degeneracy. Journal of Regional Science, Vol. 40, No. 1, 2020, pp. 89–113.

29.

Yao

Morikawa

A Study of on Integrated Intercity Travel Demand Model. Transportation Research Part A: Policy and Practice, Vol. 39, No. 4, 2005, pp. 367–381.

30.

Jin

Yao

Mode Choice Analysis in Urban Transport with Shared Battery Electric Vehicles: A Stated-Preference Case Study in Beijing, China. Transportation Research Part A: Policy and Practice, Vol. 133, 2020, pp. 95–108.

31.

Jöreskog

K. G.

Sörbom

LISREL 8: User’s Reference Guide. Scientific Software International, Chicago, 1996.

32.

Hair

J. F.

Black

Babin

Anderson

R. E.

Tatham

R. L.

Multivariate Data Analysis, 6th ed. Pearson Education Inc, Upper Saddle River, NJ, 2006.

33.

Fornell

Larcker

D. F.

Evaluating Structural Equations with Unobservable Variables and Measurement Form. Journal of Marketing Research, Vol. 18, No. 1, 1981, pp. 39–50.

34.

Adnan

Md Nordin

bin Bahruddin

M. A.

Ali

How Trust Can Drive Forward the User Acceptance to the Technology? In-Vehicle Technology for Autonomous Vehicle. Transportation Research Part A: Policy and Practice, Vol. 118, 2018, pp. 819–836. https://doi.org/10.1016/j.tra.2018.10.019.

35.

Nastjuk

Herrenkind

Marrone

Brendel

A. B.

Kolbe

L. M.

What Drives the Acceptance of Autonomous Driving? An Investigation of Acceptance Factors from an End-User’s Perspective. Technological Forecasting and Social Change, Vol. 161, 2020, p. 120319. https://doi.org/10.1016/j.techfore.2020.120319.

36.

Zhu

Chen

Zheng

Modelling the Acceptance of Fully Autonomous Vehicles: A Media-Based Perception and Adoption Model. Transportation Research Part F: Traffic Psychology and Behaviour, Vol. 73, 2020, pp. 80–91. https://doi.org/10.1016/j.trf.2020.06.004.

37.

Haboucha

C. J.

Ishaq

Shiftan

User Preferences Regarding Autonomous Vehicles. Transportation Research Part C: Emerging Technologies, Vol. 78, 2017, pp. 37–49. https://doi.org/10.1016/j.trc.2017.01.010.

38.

Huang

Hung

Y.-H.

Proctor

R. W.

Pitts

B. J.

Age is More than Just a Number: The Relationship Among Age, Non-Chronological Age Factors, Self-Perceived Driving Abilities, and Autonomous Vehicle Acceptance. Accident Analysis & Prevention, Vol. 178, 2022, p. 106850. https://doi.org/10.1016/J.AAP.2022.106850.

39.

Lavieri

P. S.

Garikapati

V. M.

Bhat

C. R.

Pendyala

R. M.

Astroza

Dias

F. F.

Modeling Individual Preferences for Ownership and Sharing of Autonomous Vehicle Technologies. Transportation Research Record: Journal of the Transportation Research Board, 2017. 2665: 1–10.

40.

Shi

Wang

Pei

The Effect of Ride Experience on Changing Opinions Toward Autonomous Vehicle Safety. Communications in Transportation Research, Vol. 1, 2021, p. 100003. https://doi.org/10.1016/j.commtr.2021.100003.

41.

Ansar

M. S.

Alsaleh

Farooq

Behavioural Modelling of Automated to Manual Control Transition in Conditionally Automated Driving. SSRN Electronic Journal, 2022. https://doi.org/10.2139/ssrn.4006056.

42.

Bierlaire

PythonBiogeme: A Short Introduction. Report TRANSP-OR 160706, Series on Biogeme. Transport and Mobility Laboratory, School of Architecture, Civil and Environmental Engineering, Ecole Polytechnique Fédérale de Lausanne, Switzerland, 2016.

43.

Kyriakidis

Happee

De Winter

J. C. F.

Public Opinion on Automated Driving: Results of an International Questionnaire among 5000 Respondents. Transportation Research Part F: Traffic Psychology and Behaviour, Vol. 32, 2015, pp. 127–140. https://doi.org/10.1016/j.trf.2015.04.014.

44.

Bansal

Kockelman

K. M.

Singh

Assessing Public Opinions of and Interest in New Vehicle Technologies: An Austin Perspective. Transportation Research Part C: Emerging Technologies, Vol. 67, 2016, pp. 1–14. https://doi.org/10.1016/j.trc.2016.01.019.

45.

Abraham

Lee

Brady

Fitzgerald

Mechler

Reimer

Coughlin

J. F.

Autonomous Vehicles and Alternatives to Driving: Trust, Preferences, and Effects of Age. Proc., Transportation Research Board 96th Annual Meeting, Transportation Research Board, Washington, D.C., 2017, pp. 8–12.

46.

Hassan

H. M.

Ferguson

M. R.

Razavi

Vrkljan

Factors That Influence Older Canadians’ Preferences for Using Autonomous Vehicle Technology: A Structural Equation Analysis. Transportation Research Record: Journal of the Transportation Research Board, 2019. 2673: 469–480. https://doi.org/10.1177/0361198118822281.

47.

Wadud

Mattioli

Fully Automated Vehicles: A Cost-Based Analysis of the Share of Ownership and Mobility Services, and Its Socio-Economic Determinants. Transportation Research Part A: Policy and Practice, Vol. 151, 2021, pp. 228–244. https://doi.org/10.1016/j.tra.2021.06.024.

48.

Yueshuai He

Jiang

Connected Automated Vehicle Impacts in Southern California Part-I: Travel Behavior and Demand Analysis. Transportation Research Part D: Transport and Environment, Vol. 109, 2022, p. 103329. https://doi.org/10.1016/j.trd.2022.103329.

49.

Benleulmi

A. Z.

Ramdani

Behavioural Intention to Use Fully Autonomous Vehicles: Instrumental, Symbolic, and Affective Motives. Transportation Research Part F: Traffic Psychology and Behaviour, Vol. 86, 2022, pp. 226–237. https://doi.org/10.1016/j.trf.2022.02.013.

50.

Zhang

Tao

Zhang

Lin

Zhang

The Roles of Initial Trust and Perceived Risk in Public’s Acceptance of Automated Vehicles. Transportation Research Part C: Emerging Technologies, Vol. 98, 2019, pp. 207–220. https://doi.org/10.1016/j.trc.2018.11.018.