Humanistic planning for urban older adults
The Elderly Acceptance of Autonomous Vehicle Services in Beijing, China
2023 年 11 巻 1 号 p. 64-84
詳細
2023 年 11 巻 1 号 p. 64-84
China's population has been ageing fast at an accelerated pace since the 21st century, while the mobility and convenience of travel for the elderly were weakened by the limited transportation options. Autonomous vehicles, an emerging mode of transport, provide flexibility for the elderly, provided that the rapid and widespread this new mode of transportation can be accepted by the elderly population. This paper plans to gain insights from the attitudes of the Beijing elderly using autonomous vehicles, perceiving the importance of different types of services. The stated preference survey was conducted over 376 respondents from various generations, which examined their perception of autonomous vehicles after their retirement. An ordered logistics model used in this paper shows that people who use emerging transportation modes (e.g., car-sharing and ride-hailing) are more likely to use autonomous vehicle services. Accessing shopping activities and medical care is the two crucial purposes for the elderly autonomous vehicle users. Additionally, the clustering and correlation analysis indicate that people with a higher willingness to use autonomous vehicles are the earliest target users of these services. However, living with children and running errands in short distances by bicycle and walk, the elderly in the early generation hardly adopts autonomous vehicle services after their retirement.
With society and economy rapid developing and the level of medical services improving, life expectancy has been extended especially for the elderly people. Population ageing is a dynamic process where the proportion of the elderly population is gradually increasing. A country or region will be defined as an “Ageing Society” when 7% of its population are people aged over 65. In recent years, more and more countries and cities are facing an ageing society (Liu, F., 2020). China will face a continuous growth in the elderly population in the next 30 years. In 2020, China's elderly population will be about 254 million, accounting for 18.1% of China's total population. The figure will reach 480 million in 2050, taking 37.8% of the total population. The aging issue in China is getting increasingly severe, while the elderly service in China is facing challengers (Yan, Gao et al., 2018). It is critical to solve the potential social problems led by the aging issue especially by identifying the demand of the elderly generation.
Table 1 shows the years when countries were defined as ageing society to different extents. Although China is ageing later and more moderately than developed countries, such as the US, Japan, and the UK, it is still noticeable that the ageing issue is increasing (Liu, F., 2020).
| Country | Early | Medium | Advanced |
|---|---|---|---|
| China | 2001 | 2025 | 2035 |
| America | 1942 | 2014 | 2030 |
| Britain | 1929 | 1975 | 2026 |
| Korea | 2000 | 2018 | 2025 |
| Japan | 1971 | 1995 | 2006 |
The emergence of ageing presented a huge challenge particularly to transportation. Social, optional and necessary are three main types for the travel activities of the elderly people (Tsai, Chen et al., 2016). With the gradual acceleration of the population ageing, the contradiction between the transport inconvenience to the elderly, and the inadequacy of existing transport services, transport facilities, related policies, and systems have become increasingly apparent. Therefore, their quality of life, well-being, and even physical and mental health may be reduced. It is inevitable for old people that their body is ageing, losing functional capacity. Nevertheless, especially in the developed countries, the proportion of the elderly with driving licences is increasing. Thus, their difficulties in driving cars will increase the likelihood of traffic accidents. On the other hand, it has been found that the routine travel of the elderly group in China relies mainly on walking and public transport. Therefore, the mobility of travel cannot be effectively guaranteed due to the limitation of walking distance, long waiting time and fixed bus routes. The inconvenience of travel will reduce the travel frequency and mobility of the elderly, especially those with poor physical health, separating the elderly from society. However, thanks to technological developments, many scholars believe that the autonomous vehicle (AV) will solve many problems related to the mobility of the elderly (Zhang, H., 2020).
The perception and acceptance of AV technology by the elderly are largely determined by whether they can benefit from it in travel. The elderly people are less receptive to new technologies than other groups, mainly because: (1) they are naturally less perceptive and are slower to accept new things than younger generations; (2) the elderly view new things more cautious and are more aware of the negative effects of new technologies. Therefore, identifying the factors that influence the perception and acceptance of the elderly to AVs, and analysing the mechanisms that influence the willingness of the elderly to use the AV services, will deepen our knowledge and understanding of the typical needs of the elderly group.
This paper examined the elderly people’s acceptance and willingness to use AVs and explored the characteristics of different user groups according to the importance of AV services. In this study, an ordered logistic model was established to understand the correlation between socioeconomic features and the acceptance of AVs. Furthermore, the k-prototype clustering technique and Spearman correlation analysis were applied in this paper for classifying user groups and identifying target group features.
With body ageing, the activity status of the elderly shows a significant deterioration (He and Ji, 2018). As an ageing society approaches, the declined mobility of the elderly will lead to a lack of social activities. In this regard, an increasing number of scholars are focusing on the travel behaviour of elderly people (Huang, J., Ma et al., 2020).
From the perspective of the lifestyles and ageing patterns of the elderly, their daily activities, especially medical care, shopping, and leisure, are important indicators of their life quality and well-being (Liu, S., 2016). Song and Wang (2018) distributed 796 valid questionnaires in Nanjing to build a structural model that not only focused on individual differences of the elderly but also considered the relationship between activity and travel. The analysis shows that the elderly mainly walks and takes public transport for running errands. The distances to the elderly taking creational activities are no more than 1.5 km. Moreover, 3km is for fitness activities and longer distances for service activities.
Some elderly people living with better health conditions and in better built-up surroundings would travel more frequently than other generations, perhaps because the majority of the elderly need to look after children plying outdoors daily (Lu, Zhou et al., 2015). For elderly people living in remote areas, the transport is too inconvenient to access to hospitals and other necessary destinations. In remote rural areas, transport service information is not up to date, which has troubled them in travel, even though they have access to significant public transport concessions. Based on the previous research using the ecological model on the travel behaviour of the elderly, perceived characteristics, attitudes, subjective well-being, and social networks are the extended and latent psychological variables having an influence on the elderly's choice of travel mode (Zhao, 2019).
The elderly in Europe and the US have higher mobility levels, preferring to travel by cars rather than by the public transport. Studies have shown that vehicle ownership and usage rates in developing countries will increase by 15-20% per year (Huang, X., Cao et al., 2012). Therefore, with the national economy improving and the population ageing, the potential for the elderly taking motorised travel will continue to increase (Li, 2021). Due to the demand for higher mobility and convenient and flexible travel, the quality of life will be further improved for the elderly people in China after their retirement. Generally, the current public transport services in China do not meet the travel needs of the elderly well (Kang and Song, 2018). With the arrival of AV services, it is expected that mobility, convenience, and safety for the elderly will be improved, although many problems are existing in the travel to the elderly. To this end, it is important to study how the elderly group perceived and accepted AVs.
Status of autonomous vehicle developmentWith the development of society and the advancement of technology, the traditional model of industrialisation followed by informatization (Figure 1) is no longer appropriate for China, and it is neither feasible nor necessary for China to follow the traditional development model. In China, two historical processes (informatization and industrialisation) are developing together (Zhang, T., Li et al., 2017).
Source: Redraw from Zhang, T., Li et al. (2017)
A study by Fagnant (Fagnant and Kockelman, 2015) suggested that AVs could be very attractive to the elderly. Harper, Hendrickson et al. (2016) indicated that AV technology could help improve the convenience for the elderly or people with disabilities in their daily travels. However, whether the elderly can benefit well from AV technology is largely dependent on their perception and acceptance of this technology. The Chinese population is more receptive to AVs than people in other countries, and most of them have a more positive attitude toward the future of AV services (Jing, Wang et al., 2021).
As people’s acceptance increases, autonomous taxis, buses and subways are beginning to land globally. Waymo (Charlton, 2019) is offering autonomous taxi services in Phoenix, Arizona, USA. In addition, AVs delivery services will also be promoted in the future. In China, developing intelligent transportation is the main goal and the new trend for future transportation. Baidu's autonomous taxi service (Jiao, 2020) and dozens of stations fully opened in Beijing. Robo-taxi designed by Xiang Dao (Zhang, Z., 2022), is the first L4-level autonomous vehicle operating platform launched in Shanghai Jiading in 2021, with safeness, joyful and accessible mobility services provided for consumers. Besides, Apollo autonomous buses are jointly developed and promoted by Jinlong Bus manufacture and Baidu in China. The autonomous bus services are now in operation in 28 cities and 39 scenery spots across China, taking on the responsibility of transport feeder at the Beijing 2022 Winter Olympic Games in Shougang Park (Zhen, 2022).
Autonomous vehicle services for the elderly groupWhat are the appropriate AV services for the elderly group that will meet daily needs and their potential requirements?
Harper, Hendrickson et al. (2016) indicated that AV technology could help improve the convenience and mobility of the elderly or people with disabilities in their daily travels. However, whether the elderly can benefit well from AV technology is mainly dependent on their perception and acceptance of this technology.
The future urban transportation infrastructure must address the enormous challenge of "not isolating the elderly". They are anxious about using AVs, so the requirements of the elderly must be fully understood and technology development and field trials must be conducted to ensure the safeness of AVs before those are deployed (Rahman, Deb et al., 2020). Therefore, training needs to be applied to the elderly to benefit from the features of AVs and adopt those into their lifestyles and driving preferences (Zandieh and Acheampong, 2021).
AVs can be used for future transport planning to solve the 'last mile’ problem for the elderly. For example, AVs could identify remote destinations, such as parks and recreational areas in the countryside, which can complement public transportation routes (Rahman, Deb et al., 2020). Kovacs analysed the prerequisites of AVs. Due to the special physical, psychological and social needs of the elderly, street design and planning policies should be considered as the factors affecting the mobility and accessibility of the elderly (Kovacs, 2020).
Autonomous vehicle technology is currently in the development stage and may not be available for public use right now. Although participants are familiar with the concept of AVs, they are unsure about many characteristics of AVs, such as the associated costs, reliability, convenience, usability, etc. Therefore, the effective way to promote AVs is to introduce how Avs work, how good these vehicles are, and what their safety features are to potential users (Lee and Hess, 2022).
One of the critical ways to increase the acceptability of AV is to apply it in public transport (e.g., buses and subways) so that users become familiar with the technology involuntarily (Gabrhel, Ježek et al., 2019). Fagnant (Fagnant and Kockelman, 2015) suggested that shared autonomous vehicles could be very attractive to the elderly. The willingness of the elderly to accept different levels of autonomous vehicles, such as semi vs. fully autonomous vehicles (SAV/FAV), is also varied. Studies have been concerned that new technologies (e.g., legal liability for accidents caused by autonomous cars) significantly reduce the willingness to use FAVs but have a much smaller effect on the willingness to use SAVs (Hassan, Ferguson et al., 2021). In addition, other research shows that the higher the automation level of an autonomous vehicle, the lower the attitudes of the elderly toward it. The results of the elderly demonstrated that SAE 2 vehicles were rated the highest, while fully autonomous vehicles (SAE 5) were rated the lowest across all attributes (Lajunen and Sullman, 2021). Rahman, Deb et al. (2020) conducted an online survey on the travel behaviour and transport needs of the elderly. Participants who can drive have a greater willingness to purchase than their counterparts, suggesting that the elderly value their ability to drive.
The willingness of them to accept AVs may vary by region. Meanwhile, the same correlation is also true for certain variables, such as gender, and other variables (e.g., race, Hispanic origin, political views) vary by culture. For example, EU and US residents may be more concerned with security and privacy (Huff Jr, DellaMaria et al., 2019). However, the Chinese are more receptive to AVs than people in many other countries, and most have a more positive attitude towards future AV services (Jing, Wang et al., 2021). Therefore, it is critical for this study to obtain the elderly’s attitudes on AV services to promote this emerging mobility service.
This study examined the perception and acceptance of elderly people toward it after the spread of AV technology in Beijing. The research structure is shown in the Figure below.
An ordered logistic model was used to construct the acceptance level of AV technology among elderly people in Beijing and explore the correlation between socio-economic attributes and the perceptions on AVs. The k-prototype clustering algorithm was used to analyse the characteristics of different AV service user groups. This study provides strategic suggestions and scientific bases to promote of AV service patterns in the future for the elderly (shown in Figure 2).
The flow of the manuscript is listed as follows: Background and literature review are discussed in Section 1 and 2. Section 3 provides a detailed description of the questionnaire design and statistical analysis according to the collected survey. Section 4 elaborates on the ordered logistic modelling and k-prototype clustering technique, examining the elderly acceptance and usage intention of AV services. Finally, the conclusion regarding policy suggestions and strategies for promoting AV services is elaborated in Section 5.
Questionnaire designTo investigate the perceptions of the elderly people toward autonomous driving mobility services after retirement, the questionnaire consisted of the following sections of respondents: personal attributes, socioeconomic attributes, and intention to use autonomous driving. To explore the intention of the elderly from various generations, the questionnaire asked the participants to envision their retirement before the stated preference survey (SP) method was employed.
On the other hand, regarding their intention to use different travel service patterns of autonomous driving, the questionnaire investigated the preferences and choices of the elderly for different service patterns by introducing the evolution context of autonomous driving and the other choices of mobility services that could be provided.
Data acquisition and general statistics Data collectionWenjuanxing, an online survey collection platform, is applied in this paper to guarantee the quality of data collection. 376 out of 482 questionnaires are collected through the survey, with an effective rate of 78%. 106 responses with incomplete progress, answering time within 1 minute and low data validity are regarded as invalid data. In addition, since it is not convenient for the elderly to receive online questionnaire survey, we adopted the offline questionnaire collection method for the elderly born prior to 1955 in this survey. The following table is a descriptive statistical analysis of collected 376 valid questionnaires. The characteristics of each variable are shown in Table 2.
| Variable | Category | Frequency | Percentage (%) |
|---|---|---|---|
| Gender | Male | 191 | 50.80 |
| Female | 185 | 49.20 | |
| Generation (Age) | Born in year of 1980-1990 | 77 | 20.48 |
| Born in year of 1970-1980 | 49 | 13.03 | |
| Born in year of 1970-1960 | 64 | 17.02 | |
| Born in year of 1960-1955 | 60 | 15.96 | |
| Born in year before 1955 | 126 | 33.51 | |
| Education | Others | 41 | 10.90 |
| High school and below | 79 | 21.01 | |
| College | 76 | 20.21 | |
| Undergraduate | 152 | 40.43 | |
| Master degree and above | 28 | 7.45 | |
| Monthly income in RMB | 2000 and below | 41 | 10.90 |
| 2000-5000 | 84 | 22.34 | |
| 5000-10,000 | 124 | 32.98 | |
| 10,000 - 20,000 | 69 | 18.35 | |
| 20,000 and above | 58 | 15.43 | |
| Household vehicle ownership | 0 | 57 | 15.16 |
| 1 | 171 | 45.48 | |
| 2 and above | 148 | 39.36 | |
| Family composition | Living with a partner | 168 | 44.68 |
| Living with children | 46 | 12.23 | |
| Living with partner and children | 110 | 29.26 | |
| Living alone | 52 | 13.83 | |
| Place of residence | Within the second ring | 18 | 4.80 |
| Second ring to fourth ring | 84 | 22.30 | |
| Fourth ring to sixth ring | 152 | 40.50 | |
| Outside the sixth ring | 122 | 32.40 | |
| The most commonly used mode of transportation | Walk | 47 | 12.50 |
| Bicycle | 47 | 12.50 | |
| Public transportation | 116 | 30.85 | |
| Emerging transport modes (car sharing, ride hailing, etc.) | 74 | 19.68 | |
| Car | 92 | 24.47 |
This survey has better theoretical representativeness of the elderly population, as the ratio of men and women was approximately comparable with 50.8% for males and 49.2% for females. However, 47.88% of the respondents obtained bachelor's degree or above, higher than the average level of Beijing, which indicates that the respondents in this study tend to be highly educated groups. In addition, as AV technology is not yet widespread at the time of survey conduction, we conducted the survey for cohorts of different generations to better represent the responses of the elderly in the future. Respondents were required to envision a retirement scenario in which AV technology fully matured and developed before they were asked to respond on the parameters associated with their retirement life. It is noticeable that the perceptions of elderly people born in different generations are obtained through the survey. In order to best picture the future situation of the respondents after their retirement life, people born in different generations they are all required to compare the people around them with similar background (including occupation type, income level, career path, etc.) for references. However, the inaccuracy due to this intuitive prediction is evitable, which could be one of the limitations of this study and need to be further addressed in the future research.
Statistical analysisFigure 3 shows the monthly income of the respondents after retirement. Most of the monthly incomes of the elderly are between RMB 5,000-10,000, and the group having an income above RMB 10,000 accounts for 34%. Income status influences elderly people's travel needs and mode choices. A stable income stimulates elderly people's demand for flexibility and mobility, increasing their acceptance of AVs services.
Figure 4 shows how many vehicles are owned by post-retirement households. It shows that most households own one or more cars. With the policy promotion and technological advancement, these vehicle ownership volumes may be converted to AV.
Figure 5 shows the family composition of the retired elderly. Nearly half of the elderly chooses to live with their partners, and up to 41% of them live with their children. It can be inferred that an independent, highly secured, and easy-to-operate mode of travel is the primary need of the elderly.
The statistics in Figure 6 show where they live after retirement. More than 70% of seniors choose to live in remote places outside the 4th ring road. That may be related to their family structure, as seniors prefer to live with their partners away from the city centre. On the other hand, as the main recreational and medical areas are clustered in the city centre, it can be inferred that the elderly will have increased travel distances and their needs may not be met by existing public transportation. That also provides a commercial chance for the application of AV services.
Figure 7 shows the statistics of standard travel modes after retirement. Taking the motor vehicle for travel by most seniors indicates that there is now a greater demand for the freedom and convenience of travel. In addition, more seniors will choose new modes of transportation except walking and taking non-motorized vehicles, which is conducive to the introduction and popularity of self-driving travel services among the elderly.
Figure 8 shows the statistics on the acceptance of AV technology by different generations of people after retirement. The generation born in 1980-1990 (now 30-40 years old) has the highest willingness to use Avs after retirement, which indicates the younger generation has a better acceptance and understanding of new transportation methods. However, the willingness of the rest four generations to accept AVs has hardly changed significantly, indicating the challenges of implementing AVs at an early stage.
An ordered logistic model is developed and analyzed by using data from 376 valid questionnaires. Gender, age, income, motor vehicle ownership, education level, family structure, place of residence, common means of travel, and purpose of travel are independent variables, while their willingness to use autonomous vehicles was used as the dependent variable in an ordered logistic analysis.
The results in Table 3 implies that there is a negative correlation between vehicle ownership after retirement and the willingness to use AVs, suggesting those who own one or more vehicles after retirement are less likely to use AV services.
The significant positive coefficients of taking bicycles, public transportation, car, and emerging transportation mode indicate that elderly people, taking most transport modes, likely have the intention on using the AVs after retirement. The highest positive relationship between emerging transportation modes and willingness to use AVs significantly shows that people who are taking emerging transportation modes at present, such as car-sharing and ride-hailing services are easier to adopt AV services in the future.
Three travel purposes - entertainment, shopping, and accessing medical care - are positively correlated with the willingness to use AVs for the elderly. Additionally, accessing medical care and shopping have the most significant coefficient values. The results show that accommodating the major travel purposes of the elderly is the critical business model for future AV services. Among them, medical care and shopping are two functions that AV service providers should focus on.
In summary, people who mainly take the four modes (bicycle, public transportation, car, and emerging transportation mode) can be potentially attracted by AV service, especially when the economically efficient, safe, and convenient AV services come to reality in the future. The elderly use AV services extensively for many purposes, and AV services are encouraged to satisfy their purpose of entertainment, shopping, and accessing medical care in the future. Furthermore, the negative correlation between vehicle ownership after retirement and the willingness to use autonomous vehicles indicates the contradiction between the desire to drive and AV services. Policymakers are encouraged to provide efficient AV services for the elderly in parallel with future policies to reduce private vehicle ownership for sustainable development transitions progressively.
| Variable | Coefficient | SE | z-value | p-value |
|---|---|---|---|---|
| Dependent variable thresholds | ||||
| Completely disagree | 3.498 | 0.812 | 4.306 | 0*** |
| Disagree | 5.021 | 0.819 | 6.128 | 0*** |
| Neutral | 7.362 | 0.868 | 8.484 | 0*** |
| Agree | 9.701 | 0.921 | 10.53 | 0*** |
| Independent variables | ||||
| Gender (male) | -0.155 | 0.203 | -0.76 | 0.447 |
| Age | -0.011 | 0.072 | -0.149 | 0.881 |
| Income | 0.161 | 0.089 | 1.801 | 0.072* |
| Vehicle ownership | -0.243 | 0.098 | -2.493 | 0.013* |
| College | 0.197 | 0.289 | 0.681 | 0.496 |
| Bachler | 0.401 | 0.265 | 1.517 | 0.129 |
| Master and above | 0.163 | 0.44 | 0.371 | 0.711 |
| Living with a partner | 0.066 | 0.313 | 0.211 | 0.833 |
| Living with children | 0.092 | 0.402 | 0.229 | 0.819 |
| Living with partners and children | 0.344 | 0.336 | 1.025 | 0.305 |
| Home location away from central | 0.234 | 0.121 | 1.934 | 0.053* |
| Bicycle | 0.87 | 0.405 | 2.147 | 0.032** |
| Public transportation | 0.958 | 0.338 | 2.838 | 0.005*** |
| Car | 0.995 | 0.361 | 2.753 | 0.006*** |
| Emerging transportation mode | 1.105 | 0.371 | 2.978 | 0.003** |
| Picking up children | 0.25 | 0.132 | 1.894 | 0.058* |
| Entertainment | 0.41 | 0.153 | 2.679 | 0.007*** |
| Shopping | 0.497 | 0.153 | 3.244 | 0.001*** |
| Access to medical care | 0.497 | 0.138 | 3.595 | 0*** |
| McFadden R-squared: 0.207 | ||||
| Cox and Snell R-squared: 0.443 | ||||
| Nagelkerke R-squared: 0.471 | ||||
Note: * p<0.1 ** p<0.05 *** p<0.01
The fitted results of the ordered logistic model only tell us whether the respective variables are significant or not, yet not tell the extent to which variable affects the willingness to use AVs. To this end, the marginal effects analysis is conducted to calculate the effect of increasing or decreasing a variable by one percent on the probability distribution of willingness to use AVs while controlling for other variables. Marginal effects are calculated by using data from 376 valid questionnaires, and the results are shown in Table 4 below.
| Dependent Variables | Coefficient | Completely disagree | Disagree | Neutral | Agree | Completely agree | |
|---|---|---|---|---|---|---|---|
|
Gender (male) |
-0.1545 | -0.0160 | 0.0792*** | 0.0098 | -0.0641 | 0.0028 | |
| (0.2091) | (0.0186) | (0.0286) | (0.0448) | (0.0490) | (0.0366) | ||
| Age | -0.0108 | 0.0021 | -0.0029 | -0.0207 | 0.0288 | -0.0080 | |
| (0.0775) | (0.0071) | (0.0076) | (0.0162) | (0.0176) | (0.0137) | ||
| College | 0.1970 | 0.0566** | -0.0744* | -0.1104* | -0.0141 | 0.0732 | |
| (0.3158) | (0.0239) | (0.0438) | (0.0621) | (0.0700) | (0.0523) | ||
| Bachler | 0.4014 | -0.0148 | -0.0253 | -0.1078* | 0.0767 | 0.0550 | |
| (0.2839) | (0.0264) | (0.0269) | (0.0581) | (0.0627) | (0.0531) | ||
| Master and above | 0.1632 | 0.0709 | -0.0573 | -0.0891 | -0.0415 | 0.0388 | |
| (0.3782) | (0.0434) | (0.0371) | (0.0866) | (0.1099) | (0.0702) | ||
| Income | 0.1606* | 0.0158 | -0.0251** | -0.0246 | -0.0232 | 0.0331** | |
| (0.0969) | (0.0102) | (0.0106) | (0.0195) | (0.0213) | (0.0147) | ||
| Vehicle ownership | -0.2432** | -0.0131 | 0.0222** | 0.0335 | 0.0051 | -0.0405** | |
| (0.1080) | (0.0110) | (0.0111) | (0.0206) | (0.0233) | (0.0185) | ||
| Living with a partner | 0.0659 | -0.0264 | -0.0161 | 0.0372 | 0.0755 | -0.0191 | |
| (0.3437) | (0.0316) | (0.0365) | (0.0710) | (0.0781) | (0.0521) | ||
| Living with children | 0.0920 | 0.0179 | -0.0137 | -0.0241 | 0.0136 | 0.0011 | |
| (0.4224) | (0.0282) | (0.0420) | (0.0981) | (0.1024) | (0.0704) | ||
| Living with partners and children | 0.3444 | -0.0524 | 0.0192 | 0.0299 | 0.0205 | 0.0388 | |
| (0.3647) | (0.0402) | (0.0322) | (0.0731) | (0.0853) | (0.0566) | ||
| Home location away from central | 0.2339* | 0.0165 | 0.0134 | -0.1006*** | 0.0448 | 0.0386* | |
| (0.1342) | (0.0128) | (0.0153) | (0.0244) | (0.0288) | (0.0216) | ||
| Bicycle | 0.8696* | -0.0302 | -0.0082 | -0.0428 | -0.0014 | 0.1318* | |
| (0.4629) | (0.0326) | (0.0486) | (0.0976) | (0.1041) | (0.0739) | ||
| Public transportation | 0.9584*** | -0.0780** | -0.0630 | 0.0301 | 0.1804** | 0.0146 | |
| (0.3192) | (0.0330) | (0.0424) | (0.0785) | (0.0808) | (0.0637) | ||
| Car | 0.9947*** | -0.0444 | -0.0881 | 0.0572 | 0.0078 | 0.1167* | |
| (0.3835) | (0.0395) | (0.0597) | (0.0828) | (0.0906) | (0.0659) | ||
| Emerging transportation mode | 1.1048*** | -0.0663* | -0.0193 | -0.0883 | 0.0753 | 0.1237** | |
| (0.3609) | (0.0397) | (0.0382) | (0.0916) | (0.0902) | (0.0618) | ||
| Picking up children | 0.2496* | -0.0067 | -0.0122 | 0.0087 | -0.0526* | 0.0622** | |
| (0.1401) | (0.0155) | (0.0125) | (0.0299) | (0.0316) | (0.0307) | ||
| Entertainment | 0.4095** | -0.0069 | -0.0329* | -0.0086 | 0.0059 | 0.0518 | |
| (0.1767) | (0.0141) | (0.0186) | (0.0323) | (0.0384) | (0.0346) | ||
| Shopping | 0.4972*** | -0.0179* | -0.0356** | 0.0045 | 0.0491 | 0.0472 | |
| (0.1890) | (0.0095) | (0.0162) | (0.0338) | (0.0362) | (0.0361) | ||
| Access to medical care | 0.4972*** | -0.0119 | -0.0074 | -0.0651** | 0.0682** | 0.0527* | |
| (0.1605) | (0.0196) | (0.0110) | (0.0298) | (0.0324) | (0.0305) | ||
| Completely disagree | 3.4982*** | ||||||
| (0.8339) | |||||||
| Disagree | 5.0209*** | ||||||
| (0.8328) | |||||||
| Neutral | 7.3620*** | ||||||
| (0.9156) | |||||||
| Agree | 9.7012*** | ||||||
| (0.9764) | |||||||
| N | 376 | ||||||
Note:a. Standard errors in parentheses; b. * p<0.1 ** p<0.05 *** p<0.01
The calculations show that income has a positive effect on the willingness to use AVs. The marginal effect explains that if an individual's income increases by one level, the likelihood of feeling “disagree” decreases by 2.51 percentage and then the likelihood of feeling “completely agree” increases by 3.31 percentage. Thus, it can be shown that as income rises in retirement, the elderly is more likely to use AVs.
Vehicle ownership, on the other hand, has a negative effect on the willingness to use AVs. The marginal effect explains that if vehicle ownership increases by one, the likelihood of feeling “disagree” increases by 2.22 percentage and then the likelihood of feeling “completely agree” decreases by 4.05 percentage. Therefore, it can be shown that an increase in vehicle ownership leads to a reduction in the willingness of the elderly to use autonomous vehicles.
Home location away from central positively affects the willingness to use AVs. The marginal effect explains that if the distance from home to downtown increases by two ring roads, the likelihood of feeling “neutral” decreases by 10.06 percentage, and then the possibility of feeling “completely agree” increases by 3.86 percentage. Therefore, it can be suggested that the elderly living far from central are more likely to use automated travel services because downtown facilities attract the elderly to travel.
The positive effect on the willingness to use AVs varies for all modes of travel, with the positive effect of emerging transportation modes on the willingness to use AVs being particularly significant. The marginal effect explains that if using a bicycle, the likelihood of feeling “completely agree” increases by 13.18 percentage. If using public transportation, the likelihood of feeling “completely disagree” decreases by 7.8 percentage and then the likelihood of feeling “agree” increases by 18.04 percentage. If using a car to travel, the likelihood of feeling “completely agree” increases by 11.67 percentage. If using emerging transportation modes, the likelihood of feeling “completely disagree” decreases by 6.63 percentage, and then the likelihood of feeling “completely agree” increases by 12.37 percentage. It can be seen that, unlike bicycle and car, taking public and emerging transportation mode not only have positive relationship with likelihood of acceptance of AVs, but also have negative relationship with likelihood of feeling “completely disagree”. This may indicate that public and emerging transport users could be the potential users of AVs especially at the early phase.
For all travel purposes, there are different degrees of positive effects on the willingness to use autonomous vehicles, with shopping and accessing medical care both having a particularly significant positive impact on the willingness to use autonomous vehicles. The marginal effect explains that if the purpose of the travel is to pick up children, then the likelihood of feeling “agree” will decrease by 5.26 percentage, and then the likelihood of feeling “completely agree" will increase by 6.22 percentage. If the purpose of the travel is entertainment, the likelihood of feeling “disagree” is reduced by 3.29 percentage. If the purpose of the travel is shopping, the likelihood of feeling “completely disagree” will be reduced by 1.79 percentage, and then the likelihood of feeling “disagree” will be reduced by 3.56 percentage. If the purpose of the travel is to access medical care, the likelihood of feeling “neutral” will decrease by 6.51 percentage, and then the likelihood of feeling “agree” will increase by 6.82 percentage and the likelihood of feeling “completely agree” will increase by 5.27 percentage. Both picking up children and access to medical care will increase the likelihood of accepting AVs, while entertainment and shopping only decrease the likelihood of rejection. This infers that picking up children and access to medical care could be the incentive of stimulating the potential demand of AVs, on the other hand, entertainment and shopping could be the supplementary service mode when more elderly people accept AVs.
Clustering analysis of autonomous vehicle servicesThis study also explored different AV services user groups and investigated how much the elderly value services, including time accuracy, in-vehicle charging service, large cabin space, adequate storage space, first-aid service, and trip tracking function. The respondents are required to rate each AV service and a scale of 1-5 represents their perceived importance (completely disagree, disagree, neutral, agree and completely agree). Based on 376 valid questionnaires, a k-prototype clustering (a combination of k-means and k modes accommodates both continuous and categorical data) analysis is conducted to classify user groups with their different attitudes on AV services.
Our user groups are clustered by the k-prototype mechanism, and the frequency and share are listed in Table 5. The average score for the perceived importance of each service was calculated. The radar diagram in Figure 9 presents the perceived importance of AV services from the four user groups (from highly valued to barely valued). On the other hand, it can be seen that there is no preference for specific AV services, which suggests that instead of providing particular attractive services for the elderly users, it would be preferable to focus R&D efforts on the essential functions, such as technology implementation, efficiency improvement, and safety issues.
| Clustering categories | Frequency | Percentage (%) |
|---|---|---|
| Cluster1 | 142 | 37.77% |
| Cluster2 | 49 | 13.03% |
| Cluster3 | 63 | 16.76% |
| Cluster4 | 122 | 32.45% |
| Total | 376 | 100% |
In order to better represent the feature of each user cluster, we renamed clusters, according to the perceived importance level of the users within each cluster (according to Figure 9). Users in Cluster 1 (Highly valued) have the highest score among all groups while Cluster 2 (Barely valued) has the least score in terms of the perceived importance. Cluster 3 (Less valued) and Cluster 4 (Moderately valued) have the score level in between. The Pearson correlation matrix helps us better understand the relationship between the characteristics of users and different clusters (Shown in Table 6). Highlighting the characteristics of different groups of users will enable AV providers to better target groups when promoting their services to the elderly population in the future.
|
Cluster1 (Highly valued) |
Cluster 2 (Barely valued) |
Cluster 3 (Less valued) |
Cluster4 (Moderately valued) |
|
|---|---|---|---|---|
| Age | -0.058 | 0.148** | 0.008 | -0.053 |
| Home away from city central | 0.053 | -0.012 | -0.044 | -0.011 |
| Willingness to use AVs | 0.342** | -0.462** | -0.117* | 0.072 |
| Gender (male) | -0.078 | 0.002 | 0.071 | 0.023 |
| Education (college) | 0.045 | 0.022 | 0.058 | -0.108* |
| Income | -0.014 | 0.062 | -0.042 | 0.004 |
| Vehicle ownership (> 1 vehicle) | 0.023 | -0.152** | -0.044 | 0.121* |
| Living alone | 0.006 | 0.028 | -0.015 | -0.014 |
| Living with a Partner | 0.05 | -0.046 | -0.059 | 0.028 |
| Living with children | -0.023 | 0.121* | 0.006 | -0.068 |
| Living with partner & children | -0.043 | -0.058 | 0.072 | 0.029 |
| Walk | -0.029 | 0.116* | 0.003 | -0.056 |
| Bicycle | -0.079 | 0.140** | 0.110* | -0.107* |
| Public transportation | 0.026 | -0.088 | -0.007 | 0.041 |
| Car | 0.054 | -0.073 | -0.073 | 0.055 |
| Emerging transportation modes | 0.001 | -0.033 | -0.007 | 0.028 |
Note: * p<0.05 ** p<0.01
The results show that users valuing all types of AV services with stronger interest and better acceptance of AVs are more likely to be categorized as Cluster1 (Highly valued). This group of people are more likely to be the target customers of future autonomous vehicle services and are more likely to be the first group to shift from other transport modes. On the other hand, college education and taking bicycles are negatively correlated with users in Cluster 4 (Moderately valued), while owning one or more vehicles has a positive correlation value. This suggests that people with one or more vehicles are more likely to be grouped into Cluster 4, and they are potential customers who are more willing to use AVs. An intuitive and bold conclusion is that vehicle owners usually travel longer distances than people taking other transport modes. The more relaxed, safer and economical trip experiences provide by AVs could be the drives that attract users in Cluster 4. However, future research needs to be conducted to consolidate this conclusion.
The elderly people classified in Cluster 2 were born in early generations, with lower interest and acceptance in using Avs, stopped owning vehicles after retirement, living with children, bicycling, and walking more frequently (Barely valued). Additionally, less willingness to use AVs and taking bicycles more frequently correlate significantly with Cluster 3 (less valued). The elderly born in the early generation are likely to be more conservative about new technologies and live with children, having a relatively lower demand for independent mobility provided by AVs. Those on foot and bicycles, more likely grouped into Cluster 2 and Cluster 3, usually travel shorter distances compare to other transport modes and not willing to use AVs, which undermines the need to apply AVs. These users are likely to be the lost customers of future AVs and are not yet ready to fully embrace AVs.
In summary, the critical strategy to promote AV services among elderly people is to increase their acceptance and attitudes on their willingness to use AVs. In this case, policymakers and AV providers should make efforts to popularize the advantage and technological advancement of AV services through commercial methods.
Result discussionThe ordered logistic regression analysis shows that people who take the four modes of transport (bicycle, public transportation, car, and emerging transport modes) are likely attracted by AV services. Elderly people used the widespread AV services for a wide range of purposes, such as entertainment, shopping and accessing medical care, and all those being encouraged to be met by AV services in the future. Furthermore, the negative correlation between car ownership in retirement and willingness to use AV suggests a tension between the desire to drive and AV services. We encourage policymakers to provide efficient autonomous vehicle services for elderly people in parallel with future policies to reduce private car ownership to achieve a sustainable transition gradually.
Also, the cluster analysis divided the population into four categories, based on the importance elderly people place on using different AV services. The study concluded that a critical strategy for promoting AV car services among elderly people is to increase their acceptance of and attitudes towards using AV. In this context, policymakers and AV providers should strive to popularize AV services' advantages and technological advances through commercial means. Moreover, the elderly who take bikes for walking distances are unlikely to use AV, while those demanding longer journeys could be a potential target for the future.
Strategies for AV business service promotion are inferred in this study. From the above survey and analysis, we can see that medical care, entertainment and shopping are the primary purposes of the elderly for travel. Since most elderly people are not accompanied by their children, the tracking and positioning functions are also essential AV services for the elderly, which can alleviate travel anxiety to a certain extent.
In addition, the quality of service is a major factor influencing whether older adults are willing to choose the autonomous vehicle. Therefore, when autonomous vehicles are manufactured, the comfort and convenience of autonomous vehicles for the elderly should also be considered, such as making the seats and facilities more suitable for the physical structure of the elderly and adding the right amount of entertainment facilities for them, to make the journey more convenient for the elderly with their sense of well-being increasing.
On the other hand, recommendations for policymakers for better implementation of AV services for elderly people are also discussed. The survey results show that nearly half of the elderly people are reluctant to use autonomous vehicles, which requires government intervention to promote AV services, especially in the early phase. Online activities, such as WeChat applets and other applications, make autonomous vehicles easier to be acknowledged.
In addition, the research also shows that the income of the elderly is not high. Many seniors believe that their income level will decrease after retirement. Therefore, the expenditure on autonomous vehicles is also an important issue that must be addressed. The government can develop appropriate incentives for the elderly, such as subsidies and pricing discounts, to encourage more people to use AV services in the future.
The results suggest that bicycle, public transport, car and emerging transport mode users are likely to be attracted by AV services. Elderly people use a wide range of AVs for a wide range of purposes, with shopping and accessing medical care being two key purposes for elderly AV users. Furthermore, the negative correlation between private car ownership and willingness to use AVs suggests a tension between the desire to drive and AV services. On the other hand, those with a higher desire to use AV are the earliest target users of AVs. However, elderly people in the early years of retirement find it difficult to adopt AVs due to living with their children and running short errands by bicycle and on foot. Elderly people who usually take bicycles for shorter-distance trips are less likely to use AV, while those who require longer distances may be potential future targets. Nearly half of seniors are reluctant to use AVs, which would require government intervention to promote AVs; the elderly people do not have high incomes and the government could create appropriate incentives for the elderly to encourage more people to use AVs in the future.
Due to the limitations of the population and the implementation survey method, the study's results may not fully reflect the relationship between all parameters and the perceptions of the elderly towards AV services. Due to some limitations in the questionnaire collection process, compared with the real level of Beijing, the respondents in this paper tend to be highly educated people. On the other hand, this paper investigated the situation of the elderly born in different generations after retirement through the SP survey. The analysis results may deviate from the actual future situation due to the limitations of the respondent's perception of their future retirement life imagined by them. Future work related to accurately picture people’s situation after retirement can be further address. With the development of the social economy and the implementation of AV technology, we must keep paying attention to the changes in the demanded characteristics of elderly people and propose travel services that better meet the actual requirements in the future.
Conceptualization, M. H., X. W., S. H., Q. Z. and T. X.; methodology, M. H., X. W. and S. H.; software, X. W. and S. H.; investigation, M. H. and T. X.; resources, X. W., S. H. and Q. Z.; data curation, M. H., X. W. and S. H.; writing—original draft preparation, M. H., X. W., S. H., Q. Z. and T. X.; writing—review and editing, M. H., X. W., S. H. and Q. Z.; supervision, M. H. and S. H. All authors have read and agreed to the published version of the manuscript.
The authors declare that they have no conflicts of interest regarding the publication of the paper.
This paper was founded by Transportation Research Base Exploratory Research Communication Program (2020BJUT-JTJDS) and Spark Program of BJUT.
