Dorian Bauer Principal UX Software Engineer • Team Lead & Technology Lead Flutter
05/05/2022 • minutes reading time
In the not too distant future, “autonomous mobility-on-demand” (AMoD) systems could provide around-the-clock, cost-effective and flexible mobility services without fixed schedules, stops or drivers, and controlled entirely via digital user interfaces.
In the research project Aperol (autonomous, person-centered organization of road traffic and digital logistics) funded by the German Federal Ministry of Digital Affairs and Transport (BMDV), we looked at the design, development and evaluation of AMoD systems - in a holistic and human-centered way. In our first Insights article on APEROL we presented the project and the project partners as well as the motivation and core questions behind the project. This article highlights the design and development process.
Scribbles
Before the journey: Identifying challenges and requirements
The development of an AMoD system requires consideration of the heterogeneous needs of diverse stakeholders – from citizens, mobility service providers, transport companies and tour planners to urban planners and municipalities. We were able to balance these needs in the project and incorporate them into the development through our collaborative, methodical approach.
The success of AMoD depends on its acceptance by people and the trust they place in it. That is why future passengers were continuously involved in the early stages of the project – for example, through various online surveys with a total of several thousand participants, or as part of a representative dialog with residents in Aachen with 89 participants. Technical, economic and organizational requirements were also defined in further stakeholder workshops. Based on these findings, we developed usage scenarios and user journey maps in order to have a common basis for discussions with our project partners.
Design process
A great ride: The design process
How can we ensure the effective, efficient and satisfying interaction of passengers and the autonomous service before, during and after their journey? This is the question we asked ourselves at the beginning of the design process.
Digital user interfaces and trust-building communication play a central role here. When designing the user interfaces, we created a detailed requirements catalog based on the user journey maps together with our project partners, which describes and evaluates the system functions and technical requirements.
The service is designed to support users throughout their AMoD journey:
User Journey Map
The development of a mobile Companion App that bundles all the services of the fictitious AMoD system for the user is essential here. We also took into account other touchpoints of future users, such as interaction in and around the vehicle, for example via passenger information displays in the vehicle interior. This also led to questions about which forms of interaction are most suitable in different situations. In addition to the “classic” graphical UI design (GUI), we took a closer look in particular at the concept of text-based conversational UIs (chatbots).
We presented the details at the ACM MobileHCI Conference 2021.
Next stop: Context-based evaluation in the Usability Lab
How does it feel to sit in a driverless shuttle? Which features in the vehicle and the app ensure trusting communication and improve the sense of security? Since there are currently not many driverless vehicles on the road, AMoD is still a somewhat “theoretical” topic with lots of open questions – a design challenge!
That is why we created a video-based simulator to realistically evaluate the concepts. The setup: We created an audiovisual simulation using audio and video recordings from real road traffic by setting up loudspeakers and projectors or screens. This relatively simple but immersive method allowed us to perform context-based prototyping and concept evaluation in the early stages of development.
Comparative evaluation of GUI and chatbot along the entire user journey in the video-based simulator.
A look under the hood: Cross-platform development with Flutter
The cross-platform framework Flutter was used to implement the APEROL frontend. Unlike other technologies, Flutter relies on its own widget library with replicas of the respective platform-native UI components. This gives Flutter full control over the rendering of the UI – regardless of the platform and its version. The tweaking of existing widgets is also quick, and custom widgets can be created by declaration with almost limitless possibilities.
This made Flutter ideal for the time-efficient as well as high-quality implementation of the APEROL Custom UI. On the backend, the app accesses a complex system of cloud-based algorithms for route planning (PSItms) and traffic simulation (MATSim) via an interface specified together with the research partners.
Developed user interfaces for on-demand test drives of the overall APEROL system in real traffic.
The home stretch: System test in real traffic with Wizard-of-Oz vehicle
After various workshops, evaluations and iterations of the apps, APEROL entered the home stretch – with a summative evaluation of the overall on-demand system in real road traffic in Aachen and Saarbrücken. The goal was to evaluate not only the user interfaces, but also the backend and route planning components in particular. A conventional vehicle (with driver) was used for this. But in order to still give the study participants the feeling of automated travel during the test drives, we implemented a Wizard-of-Oz vehicle setup.
The participants were provided in advance with the Companion App for download via Microsoft AppCenter. Our colleagues from Ergosign in Saarbrücken were able to book any two journeys on-demand on the trial days in March 2021 – for example, from home to work (in compliance with appropriate hygiene concepts due to the ongoing pandemic). In order to also examine the importance of passenger information displays in the interior, the participants received information about the course of the journey via the passenger information app during one of the two journeys, but not during the other.
The lightweight Wizard-of-Oz setup proved to be a suitable way to “simulate” AMoD in real traffic. Participants were selected when possible in advance for low susceptibility to simulator/travel sickness. The controllability of the real (on-demand) test drives was very limited (traffic volume, delays, booking parameters, etc.) compared to laboratory studies. The availability of information via in-vehicle displays seems to have a decisive positive influence on the acceptance of AMoD and the overall user experience (UX).
What is the next stop in the journey?
The findings from APEROL’s design and development process can be directly incorporated into the development of future systems. This will help shape the transportation transformation of the coming years and beyond, making it human-centered, demand-oriented and sustainable!
With the appropriate fleet size, this can lead to benefits in:
Social participation – Simplified access to mobility offers a flexible range of services, especially for people with limited mobility, and enables them to participate in society.
Sustainability – Shared travel in all-electric vehicles reduces traffic volume as well as pollutant emissions while simultaneously allowing for more travelers.
Efficiency/comfort – Driverless travel means that passengers can use their travel time flexibly – for example for relaxing or working. This results in increased efficiency and convenience.
Flexibility – Flexible booking eliminates the need for rigid schedules and stops. Users can book and customize their journeys according to their own needs.
