According to MarketsandMarkets, the projected global Robotaxi market size is expected to grow from 617 units in 2021 to 1,445,822 units by 2030 and according to Fortune Business Insights, the market value will rise from USD 1.23 Billion in 2021 to USD 108.0 Billion by 2030.
This presents a favorable opportunity to enter this thriving market. However, this also means that there will be competition from established players such as Aurora, Waymo and Zoox as well as traditional ride-hailing services such as Uber and Lyft.
Several Research indicate that people are still skeptical of Self-Driving Cars.
This research conducted by the Pew Research Center highlights what Americans think about Self-Driving Cars. For example, they mention that testing standards should be much higher for driverless cars when compared to regular vehicles.
Additionally, these are the results of a survey conducted by the Boston Consulting Group and World Economic Forum to find out consumer concerns about Self-Driving cars.
The significance of this survey conducted by Morning Consult is that on March 19, an Uber Self-Driving car crashed into a pedestrian in Arizona. This was the first known death of a pedestrian struck by a driverless vehicle on a public road. This survey shows that public opinion of self-driving cars changed significantly after this incident.
Most of the consumer concerns such as wouldn't feel safe, wanting to be in control at all times, not trusting it in mixed traffic and cybersecurity concerns are from a technological standpoint. For this project, I will primarily focus on the passenger experience that the Robotaxi offers.
We need to build trust between the Car and the Passenger. We need to focus on the passenger experience and create a positive first impression.
According to this research paper published by TU/e, we can enhance User Trust in Autonomous Vehicles by employing an intelligent system that mimics human behaviour by exhibiting these factors:
Introducing Athena AI. Going beyond a GUI, Users can also interact with Athena using a Natural Language, making it a Conversational UI. A Conversational UI is considered to be more trustworthy when compared to its GUI counterpart as per this study conducted by the University of Nottingham. The aforementioned research article from TU/e also mentions the same conclusion. Even beyond an Autonomous Vehicle, this study conducted by TU Delft concludes that users find a CUI more trustworthy than a GUI in the context of decision support.
Naturally, Athena will also come equipped with a GUI for better flexibility and accessibility (For example, a user might be socially introverted, speech or hearing impaired).
We are anthropomorphising Athena, so the passengers believe that she has some semblance of Rational Thought and Conscious Feeling as per this study conducted by The University of Nottingham.
My initial idea was to design Athena as a humanoid character and animate her facial expressions and place it on the touch interface.
But this would feel too static for Athena. She shouldn't just be something that's on a screen. I also wanted to keep the Uncanny Valley in mind when designing Athena. This made me reconsider my decision to design Athena as a humanoid agent and I wanted to find a different approach to designing this.
After giving it some thought and doing extensive research about Anthropomorpism, I had two ideas as to how we can approach this:
These are representation of both the concepts, and we can select the design that is the most plausible to implement. Both these technologies exist and are used for anthropomorphism in other use cases. One example of holograms being used is Gatebox which uses a 3D projection to display a virtual home assistant. One example of a robot companion that uses facial expressions to communicate is the Asus Zenbo.
In practice, Athena's conversational system may be used in a wide variety of contexts. Below is a table containing a list of potential conversations Athena can have with the passenger during a ride. Please note that this is not an exhaustive list.
Based on the table, I created a Voice Prototype to simulate some of the scenarios that is listed. Click the button below to interact with the Voice Prototype of Athena. This prototype was created using VoiceFlow.
A potential implementation for the conversation system can be done by combining an Intent Recognition Model and a Sentiment Analysis Model. By using this funnel, the implementation can benefit from a separation of concerns and modularity.
To simulate a model that can be used for intent recognition in a Robotaxi, I used a proxy Banking77 Dataset.
The accuracy achieved by this model is comparable to those achieved by researchers. One such example is found is on a paper titled Natural language understanding for argumentative dialogue systems in the opinion building domain where the researchers got an accuracy ranging from 86.2 to 93.9 on the Banking77 Dataset using different types of models.
To ensure that the results of sentiment analysis are tailored to our specific needs, I have developed a custom sentiment analysis model using a dataset consisting of "Positive" and "Negative" sentences within the context of our use case.
Let’s look at the options using which we can Authorise the ride. There are many different ways for doing this, each with their own pros and cons and varying levels of complexities of implementation.
After weighing the options, I have made a decision. For the rest of the project, I will be using the "QR Code" approach to unlock the vehicle and authorize the ride. This method strikes a good balance between security, ease of use, and complexity of implementation.
Before I start creating wireframes, I also wanted to visualize the end-to-end experience that a user goes through when riding in an Robotaxi booked through the app by using a Customer Journey Map. This map will guide me through the user flow when using this service, will help me make informed decisions when creating the visual design and help me identify potential pain points.
The "Authorise" step has a lot of potential to cause pain, especially when the QR code isn't being scanned properly. If this becomes a recurring problem, we will need to add secondary methods of Authentication such as Facial Recognition or OTP. The "Experience" itself is designed to bring delight, especially if it is their first time using the service. The prospect of riding in a Driverless car and talking with Intelligent System as if it were a human is sure to bring pleasure.
To start with the design of the In-Vehicle UI, I first gathered the requirements of such a system. This study conduct by Capgemini is a great starting point where they asked consumers about what features they expect from a Self-Driving car. Based on these requirements, I created an information architecture consisting of all the functionalities of the in-vehicle UI.
Based on this Information Architecture, I created two variations of wireframes, one with a vertical orientation and one with a horizontal orientation. I chose to proceed with the vertical orientation because it is better suited for displaying certain types of content, such as news articles. There is also a shift in content consumption from Desktop to Mobile Devices. We can include the option to rotate the entire screen so it's more convenient when watching videos. This behaviour will be similar to the one they exhibit when using a mobile phone. Vertical for most other activities and Horizontal when watching videos.
Based on this, I created a high-fidelity prototype of the In-Vehicle UI. While the prototype is a useful tool for demonstrating the functionality of the product, it is important to keep in mind that it may have some limitations in terms of offering the same level of flexibility as the actual product. (For example, not all buttons in the Figma Prototype is clickable.) Click the mockup below to interact with the Prototype.
To start with the design of the Mobile App, I created an information architecture based on existing ride-hailing applications like Uber and Lyft.
I then made sketches of the wireframes and once I settled with a direction, I created a high-fidelity prototype of the Mobile App. The user flow is as follows: Add a Payment Method -> Select the Pickup Point -> Select a Destination -> Confirm Ride -> Enjoy Your Ride -> Payment. Click the mockup below to interact with the Prototype.
Unfortunately, I wasn't able to reach out to a large group of participants to conduct user testing on the Voice Prototype. However, I was able to collect user feedback through the use of a SUS Test for the In-Vehicle UI.
From this, the average SUS score of all 13 respondents is 82.5. Based on Usability.gov, a SUS score above a 68 would be considered above average and anything below 68 is below average. This means that the score of 82.5 is well above average. In addition to this, I received 3 written responses of feedback from the respondents.
77.5[3]: Seems simple enough to use and understand, gives clear instructions, versatile and provides a lot of things like music and stuff. I think thats everything I need.
90[6]: It's pretty good for a prototype.
72.5[7]: The UI looks intuitive and very functional. Although the design could look more friendly to the user, it is good in the way that it looks modern.
Please note that Athena should still be considered an unfinished product. Even after we conduct the (extensive) user testing required, we will need to continuously evaluate user feedback and be ready to adjust the interface and conversation capabilities to better meet the users' needs.
Few points to keep in mind:
Remember that the intent recognition and sentiment analysis models I developed are just a proof-of-concept and require extensive research and expertise to create a more robust and accurate system. We will also need to create or collect a high-quality intent dataset to use for Athena's conversational capabilities to replace the proxy Banking77 Dataset.
Additionally, the driving and decision-making system of the Autonomous Vehicle must also be well-designed for both passenger and pedestrian safety. Assuming that the Advanced driver-assistance systems (ADAS) system is well-designed, Athena can be your value proposition that can set your service apart from the competition by emphasizing the riding experience she offers passengers.
Select a city to test the service and monitor user feedback to make the required modifications before expanding to other cities and countries. Be sure to select a city that shows more acceptance towards Autonomous Vehicles.
I hope the designs meet your expectations. In case you want to contact me for further projects, you can reach out to me at pradhyumnaag30@gmail.com.
https://www.fortunebusinessinsights.com/robo-taxi-market-103661
https://www.mdpi.com/2414-4088/2/4/62
https://dl.acm.org/doi/abs/10.1145/3342197.3344545
https://pure.tudelft.nl/ws/portalfiles/portal/120916815/3485447.3512248.pdf
https://en.wikipedia.org/wiki/Uncanny_valley
https://www.gatebox.ai/gatebox
https://www.capgemini.com/wp-content/uploads/2019/05/30min-%E2%80%93-Report-1-1.pdf