

The New York City Subway Is Using Google Pixels To Listen for Track Defects (wired.com) 23
New York City's Metropolitan Transportation Authority and Google have successfully tested technology that uses smartphone sensors to detect subway track defects, the MTA said Thursday. The four-month experiment, dubbed TrackInspect, mounted six Google Pixel phones on four A train subway cars traversing Manhattan and Queens. The phones' accelerometers, magnetometers, gyroscopes and external microphones collected 335 million sensor readings and 1,200 hours of audio data, which were processed through 200 prediction models.
The system identified 92% of defects later confirmed by human inspectors, including broken rails and loose bolts. "The goal with this [project] is to find issues before they become a major issue in terms of service," said Demetrius Crichlow, the agency's president. Following the successful trial, the MTA plans to expand to a full pilot where Google will build a production version for track inspectors.
The system identified 92% of defects later confirmed by human inspectors, including broken rails and loose bolts. "The goal with this [project] is to find issues before they become a major issue in terms of service," said Demetrius Crichlow, the agency's president. Following the successful trial, the MTA plans to expand to a full pilot where Google will build a production version for track inspectors.
Roads too? (Score:4, Interesting)
I wonder if the same tech could be used for roads. Maybe it could be even simpler, just use the accelerometer to record big jolts, and if you see a large number of vehicles getting them in the same place you know there is probably a pothole there.
Even better would be predicting failures before they happen, but just finding potholes would be a very useful start.
Re: (Score:2)
Re: (Score:3)
Not much of a point. This isn't something crowd sourced. This is something they are doing themselves. On a road you can... see the problems. There's no issue with inspection. On a subway on the other hand you need to shut down the line. Also many jolts may be the result of movements which are virtually invisible due to the rigid nature of the interface between the track and the trolly. Cars don't experience it the same way, usually only vertical (again which you can easily see).
Re: (Score:2)
Sure, but for example Google crowd sources traffic information by aggregating data from many phones travelling on the same road. If they are all going very slowly it indicates heavy traffic.
Re: (Score:2)
Oh yeah definitely if you're going a crowd sourced model then this would be the way to do it. However the problem with road maintenance is not that they don't know what needs to be maintained, it's that they don't have the budget / manpower / will to actually do anything about it. An app won't change that.
Re: (Score:2)
Using whose phones? Mounted to whose cars?
In this case, the City of New York owns the trains and the devices (phones) mounted to them. It is not a crowd-sourced "turn your personal device into a spy machine" thing.
Someone realized that all the hardware needed to perform these surveys was already available in an off-the-shelf, cheap, common device. Then they wrote some code to collect the data, mounted up the devices, compiled and verified the results.
Re: (Score:2)
Re: (Score:2)
I wonder if the same tech could be used for roads. Maybe it could be even simpler, just use the accelerometer to record big jolts, and if you see a large number of vehicles getting them in the same place you know there is probably a pothole there.
Even better would be predicting failures before they happen, but just finding potholes would be a very useful start.
They probably used to have expensive, bespoke equipment for this kind of thing, it's just been replaced with something cheaper, off the shelf that can be modified to do a similar job.
Re: (Score:2)
That's what happened in Japan with the shinkansen (bullet trains). They just retired Dr. Yellow, the special train they had for checking the tracks. Now the normal trains have equivalent equipment on them.
Cheaper is better (for wide deployment) (Score:3)
Re: (Score:2)
Exactly this! My comment would have been a copy paste. Kudos for using cheap off the shelf hardware instead of spending hundreds of millions on some dedicated hardware sensors!
Mass Market (Score:3)
Re: (Score:2)
They landed on the XBox controller, because it has the exact controls they need.
Yes I can see it now:
"Chinese ship targeted heading 270, Fire again, don't miss this time! No, fire to 268, 266, 264, 262, dammit ensign stop moving the scope!"
"Sir! I'm not touching the controller Sir! It's moving by itself slowly counter clockwise Sir!
"Mother****ing Microsoft and their cheap arse potentiometers!"
Meanwhile on the other side, translated from Mandarin:
"Admiral the Americans are constantly missing, aiming to the South."
"Yes ensign, capitalist pigs have yet to discover wonderful hall effect se
Re: (Score:2)
No need for multi-million dollars contracts just to provide custom made controllers.
How is that relevant here? You don't seem to have any idea how cheap it is to have electronic devices made now.
Re: (Score:2)
There are a lot of reasons to criticize the design of that sub, but using an XBox controller isn't one of them.
False.
I wrote a longer post [pastebin.com] but for some reason it tripped the lame filter, so you can read it on pastebin if you care.
I don't know why these B!zX clowns keep destroying usability of this site, but it's pretty pathetic.
Re: (Score:2)
Re: (Score:2)
At least in Boston there's an app you can use to buy fares and they give you a small discount.
Since all phones have amazing accelerometers now I wonder how effective it would be to have an opt-in sensor library in the fare app and gather millions of data points.
They could offer an additional discount tier too, for consideration.
Range to wifi or bluetooth beacons should be sufficient to get a good-enough average.