Hacker News new | past | comments | ask | show | jobs | submit login
Google's mapping cars discover hundreds of underground gas leaks (dallasnews.com)
245 points by state_machine on May 19, 2016 | hide | past | web | favorite | 76 comments

That's useful. The mapping cars could carry some other useful sensors. Air quality is obvious. Less obvious is RF leakage from cable TV cable, which can interfere with other spectrum users. Cable companies used to have to check this every two years, but they lobbied to not have to check it.

Google's mapping cars are the metaphorical nanites in the bloodstream we've been reading about in science fiction for decades, applied to the planet. Mingling with our other cells, traveling our veins, arteries, reporting on the general health of and providing a more complete picture of the system overall.

It would be nice if Google's streetview/mapping could be used like lint(1) to produce reports about conflicting or potentially confusing road signage and markings, so transport agencies could tune their presentation. This would all be moot after self-driving transport conquers the planet, but could be useful in the meantime.

Even after self-driving cars conquer the planet, how do you think they'll figure out that they're in a construction zone, if not road signs?

More seriously, Street View as lint for road signs would be a superb idea! It's probably not something that anyone short of Google would have the processing power to do, but it would help a great deal.

(Assuming municipalities want help, that is. There are plenty of small towns that operate as speed traps, looking to shake down drivers by posting few speed-limit signs and many traffic cops. But lint-for-road-signs could reveal when an area has unresolved road-sign issues, and warn drivers to take another route.)

Exactly, and it's sad they didn't add such additional sensors (aside from the wifi ones ;-) some time ago)

I, for one, am not keen on having nanites in my bloodstream supplied and controlled by the company who dropped from their code of conduct, and stopped using "Don't Be Evil" as their motto: http://time.com/4060575/alphabet-google-dont-be-evil/

(Top HN post in Feb 2021: "Alphabet Corp announces it's dropping support for GoogleBlood starting April 1st: A blogpost advises users to replace their blood with blood from alternative vendors and restore from backups as access to all blood functionality (not just the SmartBlood(tm) features) will cease at midnight Mar 31th. Advertisers with prepaid in-blood advertising bids will be pro-rata credited in their AdsInside account.")

> I, for one, am not keen on having nanites in my bloodcell supplied and controlled by the company who dropped from their code of conduct, and stopped using "Don't Be Evil" as their motto

I don't really know why you'd think that line would be the one thing holding them back from super villainy, but you'll be relieved to know that it's definitely still there:


I think perhaps you glossed over the part where I said "metaphorical".

Nah, I got that. I was just using your post as a setup for my GoogleBlood gag... It's probaby not as funny as I think it is… Sorry.

Just use Google Takeout.

Google Transfusion.

Damn - I should have thought of that. Well played sir!

It's only a matter of time before these streetview cars become autonomous. It's a safe bet that the self-driving project and maps teams already have a lot of interaction.

I'm a Radio Ham, and the biggest source of RF noise is arcing power lines, and those powerline ethernet adapters some people use instead of wifi.

Arcing power lines can be tracked down with with a shortwave radio and reported, but the powerline adapters are a nightmare. They're suppposed to notch out the amateur radio bands but many do not, often swamping everything from 3 to 30MHz with noise.

Are HomePlug adapters really that large a source in practice, though? I'm an amateur operator and I've spent no small amount of time messing with HomePlug adapters and various bits of equipment to measure interference. Yes, they do produce very broad noise when the network is in use (although mine, and the internet suggests many others, intentionally filter output near amateur and other in-use HF bands).

That said, I think that these devices are actually quite rare in the real world - most people don't even seem to know what they are - and in my testing the noise dissipates very quickly as I leave my house, disappearing into the background of the city after only 20-30 feet.

I think that regulators have definitely dropped the ball on HomePlug, but I also don't think that it's that much of a problem, particularly given that it's been a commercial failure. Most of the ire towards HomePlug comes from people being upset about commercial-scale BPL installations, which actually have the potential to be a major problem over a large area.

In my house, my monitors actually account for more problems I experience than the HomePlug adapters do, in my everyday radio usage. And they aren't even cheap knockoffs, they're pricey name-brand models.

It depends what country you live in. In the UK, BT Broadband, the largest ISP, hands them out for free (or used to) so they're reasonably common.

What about those adapters causes the issue? Is it the devices themselves emitting radiation? The cabling in a house?

The cabling radiates like crazy at those frequencies.

Interestingly enough, there is enough radiation that you don't really need an actual connection. Someone put one of the endpoints in a field with a generator and got a connection to another endpoint. In a sense power line adapters are just using radio over the air.

Sending data over a power line, even at low bit rates is very difficult to do reliably. So all the radiation is in a sense a feature in that it improves connectivity. Not so much a feature to those who have an actual licence to use those frequencies...

The powerline adapters works by transmitting signal over house wiring, and the wiring is effectively creating an antenna. So depending on which frequencies the adapter uses, will potentially radiate the same frequencies and the harmonics.

I'm missing something, can you help me understand?

Wouldn't ethernet cables have the same issues? Does the increased power flowing through power lines somehow amplify the signal? Is that pretty much the definition of what a power amplifier in a radio does, add juice to a signal on an antenna so it gets broadcast further?

Ethernet cables are twisted to minimize this effect. See https://en.wikipedia.org/wiki/Twisted_pair

Whoa, cool, I never thought of that!

Does the amount of wattage going through the line effect how "loud" the interference is?

Electrical wiring in your home used to pass data (as a side effect) acts as a large antenna/transmitter.

Just to clarify, as someone seeing "instead of wifi" and "acts as a large antenna" from the parent comments may come to an incorrect assumption.

In PLC (power-line communication), wiring is not meant to act as antenna. All transceivers are wired. It's an undesirable, unintentional side effect of limited (to the extent of complete lack of) cable shielding - because, well, power lines are usually meant for different application than data transfers.

Thanks for replying and pointing that out. I've corrected my comment.

So am I basically broadcasting my data to anyone nearby with a receiver?

They're encrypted (or should be), so no.

"Grow light" ballasts and solar charge controllers are big ones too.

Is there a reason they don't mess with UHF? Studying for my ham exam as we speak.

Because they don't operate in the UHF bands' frequencies. https://en.wikipedia.org/wiki/Broadband_over_power_lines#Imp... says "0-30MHZ" (that's the HF bands) "and sometimes as high as 80Mhz" (50Mhz would be 6-meters)

UHF does not propagate through power lines, the frequency is too high.

That's useful. The mapping cars could carry some other useful sensors. Air quality is obvious.

The refineries in the Houston area play all sorts of fast and loose with monitoring the chemicals they let go into the environment. If Google could measure stuff like this across wide geographical areas near the refineries, some good could come of this.

Comcast actually shut off my Internet service mid-day when I wasn't home, because of this. I was lucky that I got an appointment the very next morning (I was on-call at the time) and the tech came out, found it was between my box and their equipment, climbed up a power pole, and immediately located a wire that had the insulation chewed through by a squirrel. They seemed very proactive about shutting my node off and then fixing the problem.

On the other hand, for all I know, the problem had been there for years.

Many cable TV providers put leakage detection equipment on their trucks. RF leaks are also a point of noise ingress on the cable plant so they want to fix it to improve the cable plant.

Another hint at what is to come when Google starts organizing the world's _offline_ information.

Google Books and Google Streetview are just the tip of the iceberg.

I would be willing to bet that "massively distributed real-time data collector" was listed as a key business case by whoever pitched self-driving cars at Google.

Is the collected raw data available? Would be interesting to cross-reference what the Google cars found with 311 complaints.

Here's NYC's 311 data: https://nycopendata.socrata.com/Social-Services/311-Service-...

...I recall there being a category for people calling in gas odor, but the Socrata site seems to be slow/down at the moment.

It looks like NYC is not in scope:

  > Dallas is one of nine participating cities, including
  > Boston, Los Angeles, Chicago and Jacksonville. This program 
  > was developed with scientists from Colorado State Univ

More details on the program: https://www.edf.org/climate/methanemaps and the other cities they've mapped: https://www.edf.org/climate/methanemaps/city-snapshots

Thanks for pointing this out. I understand that EDF was the driving force behind this initiative, but the fact that the sensors were mounted on Google cars seems to have diverted attention from that fact.

There's a growing effort to monitor CO2, CH4, and CO at very fine spatial scales, and to do inverse modeling to infer emission rates ("fluxes") from the snapshots. (E.g., https://megacities.jpl.nasa.gov/portal/)

Why are these sensors not placed on garbage trucks or police cars?

Garbage truck are a really good idea. They are already going to nearly every home on a regular basis. I'm surprised the licence-plate-reading companies such as Vigilant Solutions / Digital Recognition Network aren't doing this already.


San Francisco is using ALPR on Muni buses to issue tickets to cars that are in bus-only lanes or bus zones.

But at least it's related specifically to buses and the equipment is carried by the very vehicle being impacted by the violation.


That's really interesting. In London a lot of bus lanes are enforced with fixed cameras pointing at the lanes. It means even just ducking into the lane to get past a stopped car can result in a ticket. It makes so much more sense to have the detection done by the vehicle which is (potentially) being inconvenienced by the car being in the lane.

Bureaucracy, most likely. As a researcher, attaching sensors to public vehicles is an arrangement that you'd have to make separately with each jurisdiction. With Google, there is just one jurisdiction to negotiate with.

Waste collection services are often privatized. They probably don't need the city's permission to collect environmental sensor data.

police cars don't systemically go down every street, garbage trucks, not a terrible idea, but they are inherently .. very .. dirty. I wouldn't want to be the engineer servicing sensors on them.

Also, you would probably register unreasonably high levels of all sorts of things that are being released by decomp in the garbage.

CH₄ and H₂S in particular, you would think.

We also have recycling trucks in the UK which are (where I live) just flatbed trucks with some bins on. I would imagine they would be a better platform and they also go everywhere.

So, you put the sensors on municipal vehicles. Then what? Who pays for the data upload and storage? Who pays for sensor maintenance? Who looks at it and interprets it? Who gets the benefit?

From the POV of a local government official, I'll bet this looks like a way to throw money down a rat hole and be voted out for doing so.

A local municipality can just sell/lease the space on the vehicle. The city provides 12v power and sheet metal space for bolting on a box. In return they get monthly cash payments.

Insurance companies that have GPS trackers / driving analysis boxes could add sensors and sell that data too.

Air quality sensors on garbage trucks? Let me guess, it would be biased :)

Implying air quality sensors placed on any road vehicle aren't biased? While the majority of road vehicles still burn fossil fuels, it would be stupid to ignore the impact the fumes have on the data. It's important to remember that they are measuring the air quality a meter or two off the surface of the road network, NOT the region as a whole.

However if all you are looking for is major gas leaks, even the data gathered from garbage trucks aught to be accurate enough. Gas from leaking pipes will be far more concentrated than that produced by garbage trucks.

Unfortunately, I have not read the article due to the site apparently dying to the Hacker News hug of death.

They could switch from Subarus to Priuses. :)

It is kind of interesting they chose Subarus instead of Priuses. I wonder what sort of data/calculation they have saying some mid-sized 4WD ICE-only vehicle is the best for street view driving.

If I had to guess, I'd guess that, similar to their datacenter setups in some cases, they chose the cheaper, more easily replaceable option over a more efficient but {constrained, expensive} option.

What did they say when you suggested this to your local sanitation and police workers?

Or USPS vehicles, for that matter....

Also public transport. In my city they already have gps and network connection and there is a website where you can look up where they are.

Or USPS delivery jeeps/cars/trucks.

This case sounded a bit experimental? Maybe they could try those vehicles next now that they know that the sensors work.

Slightly off topic, but one idea I had was for Google to share road usage stats with cities and businesses. Right now cities put out an air tube on the road that counts the cars. Google could provide much more insight given they already can tell you road congestion. They could give it away for free to a city and/or make businesses pay for it.

I think this is a service Strava offers, but for bicycles.

I wonder if accelerometer data would be useful to determine road quality? It may help cash strapped cities like here in Toledo prioritize the bumpiest roads to fix.

This strikes me as an excellent IoT application, gas sensors are cheap and simple SMS texting could send a message from an IoT device that it detected gas and its GPS co-ordinates.

> gas sensors are cheap

I wanted to buy one to check my pipes for leakage, but the only inexpensive thing I could find was a go/no-go sensor, and nothing something with an actual reading.

This one (http://www.seeedstudio.com/depot/Grove-Gas-SensorMQ2-p-937.h...) is $7, Its sensitivity can be adjusted by the pot, so putting a digital potentiometer on the board and driving it remotely should give you some "range". It isn't designed as a precision instrument but in an IoT situation just watching sensors detect something over time, correlated with wind speed and direction, should give useful information for the gas company at least.

These guys (http://lmsnt.com/applications/applicationideas/ch4/) will sell you a sensor development kit. The commercial instruments are all several hundred dollars but my experience with specialty tools is that they have high prices to support high margins because they don't sell in very high volumes.

There are a few factors.

The gas sensors you can afford as a (poor) hobbyist tend to be resistive/heater devices. You heat up a catalyst and monitor its resistance which varies when a suitable gas is present. The downside is that typically they're sensitive to all kinds of gases - e.g. methane, ethanol other volatile organic compounds and so on. Either you can use them in a single-gas environment, or where you would only expect one polluting gas, or you can have a general threshold. Cheap and cheerful and nice for IoT home "air quality" sensors, but useless for real measurements.

Most of the very expensive sensors - like the one you linked - have much higher specificity. That methane detector is optical and is tuned precisely for the IR absorption band of the gas. Optics and filters are expensive, plus the driving circuitry and probably NIST-traceable calibration if you want it. EDIT: It also uses a fairly unusual LED wavelength which adds to the price (InGaS = money).

Other sensors, like Oxygen, have shelf lives of a few years and need to be replaced even though they're horribly expensive ($200 or so).

A filtered methane sensor is about $50 (Figaro make a lot of nice gas sensors): http://microcontrollershop.com/product_info.php?cPath=301_42...

AMS also make some nice inexpensive general VOC sensors like this one: http://www.digikey.co.uk/product-detail/en/ams/AS-MLV-P2/AS-...

From the article:

residents paid as much as $1.5 billion extra between 2000 to 2011 for gas lost to leaks.

Is anyone able to calculate approximate how much gas this is in volume or weight? 1.5 billion divided by residential gas price per unit?

Call it $10 per thousand cubic feet.


So 150 million units. Switching to the numbers I'm seeing quoted for pipelines, 150 billion cubic feet.

This pdf lists many pipelines pushing 1 billion cubic feet or more a day:


So at most a couple months of normal usage.

Check out this chart from the Energy Information Administration: https://www.eia.gov/dnav/ng/hist/n3010us3m.htm

If you only want 1 sigfig, I'd say the price is about $10 per thousand cubic feet over that timespan. So that's 150 billion cubic feet? You could calc a mass if you knew the composition, or a weight if you knew the pressure.

And all they had to sacrifice was personal privacy!

Like all great futuristic advances we've seen in the last two decades that REALLY put us into the "we're living in the future" category, there's a plus and negative side.

I love that I can wear a watch that doubles as a phone, health monitor, gps/mapping, mp3 player, etc.

The flip side of there being a loss of privacy would both be technological (weak or no security in the device itself or the apps and services it uses) and political (government saying "give me all the data").

This cry of "but my privacy!" is nothing more than a useless complaint, and does nothing to address the problem. We as citizens and consumers of such great tech NEED to demand better device security as well as political reform. Changes in how things are done and why.

So be picky about which tech you buy and politicians you vote. Be picky with your time, money, votes.

Don't, however, waste time on comments that do nothing to address the problem. That includes providing no information to further advance your ideology: privacy. No product endorsements to sway us into purchasing or supporting something which WILL keep your private info out of nosy third-party hands, or info about politicians or policy we should be aware of?

That would be a pointless comment when you obviously care about something.

I don't carry a smartphone, I don't have a Facebook account, and I use NoScript, so I have a certain amount of privacy online; but there's not much expectation of privacy in public places. (Not zero, evidently -- Google blurs out faces on Street View, for example -- but not much.)

And in this case, Google's not revealing personal data at all, they're finding and reporting industrial problems; their wide-ranging, unaccountable cyberstalking powers are very creepy, but this time they're being used for good.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact