Carnegie Mellon — IxD studio
Steerra
A speculative interaction-design case study reimagining the scooter as a semi-autonomous vehicle — designed to earn rider trust before the system ever takes over.
Summary
Role
Designer working with 2 other designers
Timeline
January — May 2024
Responsibility
I conducted research, ideated, created flows, and designed screens for the final product.
Problem
How can we create a semi-autonomous scooter that brings value to the market that:
Distinct from other transportation.
Create value that doesn't exist with other forms of transportation.
Build on scooters' existing value.
Retain the benefits current scooters already provide.
Solution
Prioritize safety by surfacing as much status information as possible during autonomous driving.
Provide instantaneous manual takeover
Communicate intention to other vehicles without sacrificing safety
Change the form factor of a traditional scooter without sacrificing compactness and convenience
Research
Why riders pick scooters in the first place.
Conducting guerrilla research with 9 current scooter users, we sought to discover why people preferred to use their scooters over other modes of transportation.
Convenience.
Scooters are popular for quick commutes, especially among students. They're easy to start and park.
Portability.
Scooters are quite easy to transport and store.
Effortless.
Riding a scooter is not physically exhausting.
Safety is a huge concern
Riders feel exposed and uncommunicative on the road.
Interviews revealed major safety concerns, highlighting the need to prioritize rider trust and confidence — especially when designing an autonomous driving feature.
Safety
Riders felt unsafe in traffic, often squeezed between cars. Though electric scooters are restricted to roads, many ride on sidewalks to avoid danger.
Lack of road communication
Electric scooters lack safe ways to communicate with cars — hand signals require releasing a handlebar, reducing stability and safety.
Trust is everything
One very eye-opening and insightful question we asked helped lead our design for our semi-autonomous driving system.
Distrust runs deep
Riders don't trust autonomous driving systems. A lack of control is a contributing factor.
How might we…
With safety, and more specifically trust as our top priority, we focused on designing a semi-autonomous scooter while preserving the value of manual riding.
develop a semi-autonomous scooter system that fosters rider trust, enables on-road communication, and preserves the core benefits of scooters?
What even is a scooter?
We wanted to take a step back and look at what a scooter is defined as.
Since traditional scooters require two hands for balance, using a screen compromises stability. We addressed this by redesigning the scooter with a seat and steering wheel — and made sure it still folded down for the convenience riders love.
Screen flow
With a screen in the steering wheel, we prioritized key displays.
Manual mode limits riders to default screens for focus, while automatic mode allows swiping through more information to build trust — especially the Automatic Driving Visualization to see what the scooter sees.
Improvements
Refining the design after rider feedback.
Following our ideation of what our scooter could look like and the decision flow, we designed our screens to determine what could be improved.
Automatic Driving Visualization
Riders wanted to immediately see info on what the system visualized when automatically driving in order to trust it and make themselves feel more safe.
Countdown for system manual handoff
Originally, system handoffs (scooter prompts switch from automatic driving to manual) were immediate, but riders needed time to prepare. We added a countdown to give time to be aware for smoother takeovers.
Remove switching to automatic mode on screen
We disabled the ability to switch to Autonomous Mode from Manual Mode using a button on the screen. Riders sometimes took their hands off the steering wheel during Manual driving to press the button, which is dangerous.
Metrics
What I would have measured.
We didn't have enough time to fully build out our final scooter prototype and run a Wizard of Oz test, but if given the opportunity, I would focus on evaluating:
Trip completion efficiency.
Measure the total time required for a user to start, ride, and complete a trip in autonomous mode.
Perceived safety and trust.
Gather user ratings (1–10 scale) about how safe and trustworthy the scooter feels during automatic driving.
Manual override responsiveness.
Measure how quickly and easily a rider can switch from autonomous to manual control, including transfer reaction time and user-rated ease of takeover.
Reflections & learnings
Iterate, test, learn fast.
We built cardboard models of our scooters quickly to test out concepts fast — not the prettiest, but it's the most efficient way to move forward in the process.
Challenge the norm.
I am really proud of how we questioned the definition of a scooter and redefined and redesigned a scooter different from what a traditional one looks like. Not only did we reframe the design brief, but we reframed the core definition of the product.