Social desirability and mobility impacts of early forms of automated vehicles

Abstract

Double degree thesis: Part A: Civil Engineering Part B: Science Communication The first forms of automated vehicles (level 1 and 2) are already available at dealers, and next levels are being developed at this moment (level 3 and up). Literature indicates two development paths for automated vehicles: an autonomous and a cooperative path. Autonomous vehicles only monitor the driving environment, whereas cooperative vehicles also communicate with other vehicles or roadside systems. This thesis consists of two parts: one (part A) researching the mobility impact of these two development paths, the second (part B) develops a method to include the public in decision making around automated vehicles. Part A: Modelling the mobility impacts of automated vehicles Governments are eager to know the impacts that automated vehicles have on mobility. Investment plans and policies can be made with this information. Current macroscopic models that assess the large-scale impacts of automated vehicles are complex, unsuitable for explorations with many uncertainties and are not able to simulate multiple vehicle types. This thesis aims to explore the impacts of early forms of automated vehicles (level 1, 2 and 3) on mobility. To cope with this problem a System Dynamics model (SD-model) is built. This model is based on the structure of the ScenarioExplorer, a model developed by TNO in the 1990s. The SD-model is strongly explorative and does not make use of an explicit road network. The goal of this model is to capture the most important effects of automated vehicles, but not to go into all the details. As the structure is simple and the run time is short, the model can be used to assess different scenarios. In this model the road capacity, value of time and fuel economy effects of automated vehicles are researched. The different levels of automated vehicles are modelled as different user classes in the mode choice, time of day choice and the assignment. This is novel for modelling automated vehicles on a large scale. In the assignment PCU factors depended on the penetration rate are used per vehicle automation class. This PCU makes it possible to translate results of microsimulations easily to large scale models and to simulate mixed traffic. The SD-model is compared to three macroscopic models and historic data and shows similar results. In addition, other tests point out that the model is suitable for explorative studies. Simulations with the SD-model show that due to the benefits automated vehicles bring, they will lead to extra car traffic in all researched scenarios. In the cooperative development path, the travel times on characteristic relations will roughly stay the same due to capacity benefits. In the autonomous development path, the average speeds drop due to less capacity benefits. The model shows that early forms of automated vehicles will not reduce congestion and in most scenarios have a negative effect on mobility. The only benefits early forms of automated vehicles entail are for the drivers, but not for mobility as a whole. Governments should therefore invest in other measures to stimulate the mobility. Due to the increase in car traffic, more emissions are expected. Part B: A more responsible innovation through the use of a constructive dialogue Societal impacts of automated vehicles can be large, not only on mobility, but also on safety, privacy or security. Complicating aspect is that automated vehicles both influence the living environment of the consumers and other road users. Literature indicates that at this moment the public (both user and other road users) are important stakeholders, but are not enough involved in the automated vehicle innovation. Due to this, and other flaws, the automated vehicle innovation cannot be called a responsible innovation. not involving the public constitutes the risk of neglecting their fundamental ethical principles, as their opinions remain unheard. This research aims to develop a method to involve important actors and to translate their ethical principles into starting points for a design of future automated vehicles. The values of four important actor groups (the government, manufacturers, consumers and non-consumers) are investigated. The method aims to be a constructive dialogue method. The value profiles created from the questionnaire show that opinions of the various actors differ. All actors agree that safety is the most important value. Differences are that the government and non-consumer value traffic flow, whereas the car manufacturers value spending time differently and self-determination for the driver. The cooperative path therefore seems attractive for the government and the non-users, whereas the car manufacturers are most likely to be in favour of the autonomous path. The survey shows no preference for one of the two paths from the consumer. The value profiles created from the questionnaire show that opinions of the actors differ. All actors agree that safety is the most important value. Differences are that the government and non-consumer value traffic flow, whereas the car manufacturers value spending time differently and self-determination for the driver. The cooperative path therefore seems attractive for the government and the non-users, whereas the car manufacturers are in favour of the autonomous path. The survey shows no preference for one of the two paths for the consumer. To create a common value profile a dialogue is needed, this is done in workshops. Tests with the constructive dialogue workshops show promising results: tensions in values become clear and the students reach consensus in the workshop. This is empirical evidence for what Van de Poel (2013) describes in his paper on specifying values to design requirements. The set-up seems to be a way to involve the different actors. This method is therefore a step towards a more responsible innovation for automated vehicles. Another promising aspect of the method is that the new ideas which are not mentioned in literature on self-driving cars arise in the sessions. This research contributes to a more responsible innovation as stakeholders are involved. Still, other important steps have to be taken. The method which is developed in this thesis should be used by manufacturers to give input to future designs or by governments for policies. Future research should focus on the validation of the workshops and the embedding of the method.