In October, I was selected from hundreds of applicants to spend a day at IDEO CoLab to participate in their makeathon with fifty others. It was a day full of forming new friendships, post-it-ing, and sprinting to design new solutions to difficult problems in an emerging space. We were assigned a team of 4 other people with backgrounds in design, tech, and engineering from all stages of life.
How might we crowdsource real driving data to improve autonomous vehicle systems?
Ideas listed included LIDAR and camera data in cars.
Our team went through the entire design process, from understanding the prompt, to defining the user journey, creating how might we statements, ideating, and then finally designing and prototyping. Design is democratic, so we spent a lot of time discussing ideas and voting to narrow down the scope to create a feasible product.
There were originally 4 types of potential consumers we wanted to understand:
- the commercial driver: fedex, ups, truck drivers.
- the commercial driver: rideshare drivers
- the personal driver: driving for pleasure & work
- the autonomous vehicle (AV) companies that needed data
We had so many ideas for data points and platforms, from building sensors into headlights to incentivizing drivers to share data by helping them find lost pets (!). Ultimately we narrowed down to helping commercial rideshare drivers and AV companies by creating a marketplace platform for drivers to integrate data from their driving to sell to companies securely through the blockchain.
THE DATA CONSUMER:
Meet nuTonomy, an AV startup.
nuTonomy is helping develop the next generation of autonomous vehicles, but they need data to train their autonomous vehicle systems.
Quantity of data: Companies like nuTonomy do not have the car fleets to amass the quantity of data needed to train AV systems
Quality of data: Engineers need data from a specialized suite of sensors to train AV systems.
Communicate with the data gatherer: Companies like nuTonomy need to communicate with data gatherers obtain data on specific maps.
THE DATA GATHERER:
Meet Ralph, a Lyft driver.
Ralph is trying to make the most of his time working as a Lyft driver in the Bay Area
Make more revenue: Ralph is trying to maximize his revenue from working as a Lyft driver.
Locate riders: Ralph needs help locating passengers in congested urban areas and at night.
Improve safety and convenience: Safety for Ralph, his passengers, and other pedestrians.
Carl is a car add-on that can create and collect all type of data, including utilizing a built-in LIDAR system that detects surroundings and is the key technology in autonomous vehicles.
For drivers, Carl will enhance the ridesharing experience by helping drivers like Ralph find passengers, earning extra revenue, and create a safer experience for drivers and riders. The built-in screen can serve a variety of purposes.
For AV companies, Carl will allow AV companies to amass high quality, specific data for training, and spend less time and money collecting or purchasing data from other organizations.