The New York Metro uses Google Pixels to listen to track defects

Between September and January, six Google Pixel Smartphones got free of free rides on four New York Subway cars. In particular, they took the train A, as he spoke 32 miles between the northern end of Manhattan and the southern Queens.

The phones were not squeezing or homeless, and extremely sharp eyes could understand, as they were inside the plastic enclosures and were fixed by brackets to the underside and the interior of the cars. While people in cars used their smartphones to write emails or to scroll Instagram or to examine Roblox, metro operators used sensors on these phones – Accelerometers, Magnetometers and Gyroscopes, and for those attached to the exteriors of the vehicles, additional external external micro -heads

Phones were part of a brief experiment from New York Metropolitan transport organ and Google Whether the cheap, especially the technology outside the shelf, can supplement the work of checking the agency. (S)Public Sector of GoogleThe department that did the job did not load the MTA for this initial experiment. Today, inspections are carried out by human inspectors, which together run all 665 miles from the New York subway, the eyes peel off for problems such as broken rails, destroyed signals and water damage. Three times a year of specialized vehicles, loaded with “train geometry cars”, also capture and upload more complex data on the condition of the city’s railway infrastructure.

New York transit work with experimental technology that Google calls Trackinsspect suggests that audio, vibrations and location data collected relatively cheaply and are used for training artificial intelligence The forecasting models can complement this check. This can direct people to suspicious rattles, bangs or screams, which suggests what types of tools they will need to make repairs before they get there. During the four-month project, the technology was able to identify 92 percent of defects later indicated by the human track inspectors, MTA says.

After all, the technology can become a “way we can minimize the amount of work done to identify these defects, and to direct inspectors in the right direction, so they can take the time instead of identifying and go directly there and do the job,” says Demetrius Crushlet. In the future, MTA hopes to create a “modernized” system that automatically identifies and organizes adjustments to the song’s problems.

For 3.7 million everyday system riders, the capture of defects, before problems, may be the difference between getting to work or school on time and sinking into unexpected delays.

“The goal of this (project) is to find problems before they become a major problem with the service,” says Krichlow. Cooperation with Google will already expand to a full pilot project, says MTA, where Google will build a production version of the technology and put it in the hands of the track inspectors themselves.

Inspectors

The Google experiment is part of an AI -enabled AI harvest that transit agencies are just beginning to use to supplement their typical checks, says Brian Music, Vice President of Transit and Consulting WSP. While New York is unique in using “harmonics” – angular and vibration – to determine problems, others have installed small sensors or song cameras that make automated measurements and non -compliance in the flag when they appear. The technology is activated not only by the progress in machine learning, but also the cheaper and smaller batteries and processors.

However, US regulators require regular human check and support, and Poston says they do not expect these rules to disappear soon. “While technology cannot be specific and precise, you will always need this human interaction,” he says.