MIT Students Launch “Curb to Car” App to Help Blind Users Locate Rideshares

A team from MIT Assistive Technology Club has released a new iPhone app designed to help blind users identify and locate their rideshare vehicles. Speaking on Double Tap, student developer Tage Mehta confirmed the app, called “Curb to Car,” is now available and aims to solve a common challenge at busy pickup points such as airports.

The app uses a smartphone camera and computer vision to scan for vehicles matching a user’s ride details. Mehta explained the goal was to make the process “seamless,” allowing users to “hold your camera up and it’ll start talking to you, tell you where it is, how to get there.”

The system works by copying vehicle details—such as make and colour—from apps like Uber into Curb to Car. It then continuously scans the environment, identifies nearby cars, and provides audio guidance including direction and distance. Blind advisor to the product Scott said the app “will walk you right to the car,” using audio cues and spatial feedback.

Development presented technical challenges. Mehta noted that identifying specific vehicle models required large datasets that were not publicly available. The team instead partnered with an external service to improve recognition accuracy. “There just isn’t enough publicly available data to train a model like that,” he said.

The app combines on-device detection with cloud processing. While initial image recognition happens locally, images are sent to a third-party service for detailed identification, meaning an internet connection is required.

The tool also adapts to available hardware. Devices with dual cameras or LiDAR provide more accurate depth information, while single-camera phones still offer directional guidance.

Currently iPhone-only, the team said an Android version would likely require additional development support. They are also exploring partnerships with assistive technology companies and potentially rideshare platforms to ensure long-term sustainability.

The app addresses a long-standing accessibility gap in rideshare services, particularly in complex environments with multiple vehicles. By offering continuous, real-time guidance, it provides an alternative to visual identification and reduces reliance on external assistance.

Resources

• MIT Assistive Technology Club https://assistivetech.mit.edu