Self-driving car tests collect too much data

Self-driving car tests collect too much data

Photo: Allen J. Cockroaches/Los Angeles Times (Getty Images)

The affair of self-driving cars is not really a car company. It’s not about mechanics, power, handling and borderline understeer, but about software — trying to build the brain of a car to make better decisions than a person.

For now, it’s a pipe dream, but it’s a pipe dream built on data. Data from every car in the fleet, analyzing everything around it thousands of times per second. All of this information helps build algorithms, to make autonomous driving smarter, but it comes at a very real cost: Storing, maintaining and using this data is not cheap.

Wired dives into the world of AV data storage and found that companies are reducing their tendencies to hoard information. As Fleets Grow and the price of cloud storage continues to rise, autonomy developers no longer see the benefit of storing every data point from every sensor, always and forever.

This is a major change for an industry that ruining infotainment systems with its data collection trends, but necessary for companies that want to reduce costs. Wired cites a statistic saying that in just one hour of operation, a Waymo Jaguar generates 1,100 gigabytes of information. — which would cost the company $22 per month to store on Amazon’s AWS infrastructure. Not much in a vacuum, but think about how many cars Waymo has and how many hours it’s been running. A day of operation would generate large amounts of data, the price of which would continue to increase.

All in all, companies that collect less data are a net good. This may have a detrimental effect on the Autonomous Intelligence upgrade rate, but it will help keep more businesses in the game.


Leave a Reply

Your email address will not be published. Required fields are marked *