Samsara Develops AI Model to Detect and Track Pothole Deterioration

Fleet management company Samsara has developed an AI model that detects multiple types of potholes and determines how quickly they are deteriorating. The San Francisco-based company has spent the last decade equipping millions of trucks with cameras primarily for driver monitoring, theft prevention and liability claims.

It trained the model on that collected data. Samsara said its trucks are far more prevalent than robotaxi fleets currently operating on public roads. The company stated this allows it to gather more data including repeat observations from the same locations that show changes in potholes over time.

It believes the information will help cities maintain roads more effectively. The company announced Tuesday that it has multiple cities under contract for the service called Ground Intelligence. Chicago is coming on as a new customer. The offering is the first in a series of insights and data points that will be provided through the platform.

Intelligence works as a dashboard that proactively populates warnings on a map showing developing potholes and other potential problems. It also allows cities to pull anonymized footage from vehicle cameras to confirm citizen reports of issues such as downed street signs or clogged sewers.

A company spokesperson said the system turns a reactive process into a proactive one. Typically cities dispatch workers or review hundreds of 311 calls to identify infrastructure problems. Samsara's pitch centers on the large number of commercial trucks and vans already using its cameras to deliver targeted information quickly.

The spokesperson added that this approach enables planning to address multiple potholes in one area during a single outing.