UberSELF - Design Study for Uber self-driving cars


Design Study for Uber's Self-driving Cars

Disclosure: This is not an actual product of Uber. This is a personal design study proposing a potential feature for Uber self driving cars. 

Description Personal Design Study Created in 2/27/2017

Software Sketch Principle


According to a study by University of Michigan's Transportation Research Institute, there was a continuous decrease in the percentage of people with a driver's license for people in the US age 16 to 44 from 1983 to 2014. On the other hand, there was an increase in people at the older end of the age spectrum (70 years old or older) getting their driver's licenses. People's reflexes and vision inevitably worsen as they grow old. Self-driving cars could allow people of all ages to easily get around without the need to drive. 

But how will people start adopting self driving cars? 

Daimler(parent company of Mercedes-Benz) has partnered with Uber, Lyft with General Motors. Waymo has teamed up with Fiat Chrysler and there are reports that the two will launch a robot taxi service by the end of 2017. As a leading platform in ride-sharing, Uber could potentially serve as people’s main mode of transportation. I wanted to imagine how people’s riding experiences could be enhanced if Uber was to launch their self-driving cars now.


Key Questions

Uber has recently started testing it's vehicles in Arizona (see more details in this article). Right now, a passenger could ride UberX in a self-driving car, accompanied by 2 Uber engineers for safety reasons. In the future, when you are riding alone, how would this experience be fundamentally different from a normal Uber ride? 


User Flow

I decided to focus on the riding experience of the following flow.


1. Call UberSELF

Although I'm not sure how the pricing would exactly work for uber self-driving cars, I decided to include uberSELF as part of the Economy group based on the fact that current Uber self driving cars in Arizona are piloted as part of UberX. I speculate that once self-driving cars become more prevalent, uberSELF will also have options for POOL. Click here to see another design I explored.


Detail Views


2. Find Car

Normally, if you can't find your Uber car, you can call your driver or the driver can call you. Since there is no one you can call in a self-driving car, I thought of 2 ways that could help people find their car. 


Car has arrived but I can't find it.


Option 1 - Leveraging Uber Beacons to find car with blinking lights

Last year, Uber has piloted the use of beacons in Seattle. The beacons is a device that goes on a driver's windshield and uses color-pairing technology to help drivers and riders more quickly connect at night, particularly at crowded venues. Currently, these beacons are only available in selected cities, but Uber has announced to use it as a new standard for the Uber experience in 2017.


In line with the expansion of Uber beacons in 2017, UberSELF could incorporate beacons to help people easily recognize their cars.


Press button to make beacon blink


Change color of the beacon



Option 2 - Identify the car's surroundings

Currently, the Uber self-driving car has 20 cameras capturing views of the front, sides and rear. 

Sometimes, it's difficult to know what side of the road the car is located, especially if the car is parked at a crossroad. What if we could leverage the existing cameras in the cars to help riders identify the cars exact location?


See photos of the car's surroundings

Created in Principle


I still can't find my car.

In the case that the person still can't locate the car, he/she can call an uber representative for help, order a new ride or cancel ride completely.
Click here to see another design I explored.


3. Confirm Rider

I got in the car but how will the car know I'm the right passenger?

Normally, when I get into an Uber, the driver asks "Are you Jasmine?" or I ask "Is this ride for Jasmine?" to confirm that I've got in the right car. Since there is no one to confirm with, I speculate that there would be a virtual key of some sort (using bluetooth tethering or RFID keyfob) to open the car. Hence, as you approach the car, the car would automatically recognize the right passenger. However, depending on the bluetooth signal or sensor malfunction, I believe QR codes could be utilized for double confirmation. In the case that you're ordering a ride for a friend, you could also send this QR code so the person riding could access the car.


4. Ride

Have full control over your ride.

Ever experienced that awkward moment when you want to ask the driver to change their terrible music but you can't? Well, in self-driving cars, you can fully enjoy your ride in your comfort zone. You can easily change the temperature, music and radio from the app. Click here to see another design I explored.

How could my ride feel more human?

Sometimes, we have great conversations in the car that allow us to connect with the driver and make our overall ride feel more meaningful. What if I can talk to the car instead of having the car passively respond to my controls? To simulate a conversation, I used the IBM Watson Conversation API to map out a conversation that the car and riders could potentially have. 

Note: The conversation below would be verbally communicated, not through text. 

Conversation Map



5. Rate

How did I feel about my ride?

In theory, self-driving cars could solve traffic issues by eliminating human reaction time issues and communicating with other cars to maybe even remove the need for traffic signals. However, one slight miscalculation in milliseconds in a self-driving car could spark traffic jams that can reach for miles. Since the "timeliness" of self-driving cars will be critical in the future, I decided to include that as an element.


During this design study, I had a chance to assess the appeal and challenges of self driving cars. Here is a summary based on what applies now and what will apply in the future when self-driving cars are a common mode of transportation.