New Navigation App 'Live Roads' Promises 1.5m-Accuracy With Standard Cellphone Hardware (arstechnica.com) 80
Jonathan M. Gitlin from Ars Technica reviews a new navigation app called Live Roads, which promises 1.5m-accuracy via your current smartphone without the need of any extra hardware. In a nutshell, the app provides more accurate mapping/navigation than what's currently available via Google Maps or Apple Maps, but it's still not quite as accurate as a true "HD map." HD maps are accurate to within a centimeter or two and are usually made by a combination of traditional surveying and lidar scanning. Here's an excerpt from the report: A few weeks after talking with the company, I was delivered a Samsung S7 loaded with Live Roads. I'll be honest: I'm not that familiar with Android, and this isn't really a review of the app. I used it enough to check that it does what it claims, but I didn't use it as my sole method of navigation. However, this brief bit of user-testing did let me check out the claims in that email. I don't think I'd equate the app with the HD maps that autonomous vehicles will need. For one thing it's readable by a human being; for another it's not quite that accurate. But the spatial resolution was indeed better than it should be on a consumer phone, and Live Roads was able to locate me down to a specific lane on a multi-lane road. Various navigation apps give you lane-specific instructions -- for instance, telling you to stay in the middle two lanes if you're approaching a complicated intersection. Where Live Roads differs is that it can also tell which lane you're actually in. Whether this is enough of a feature to build a business model around is an open question; I'm quite happy using Google Maps on iOS, with occasional forays into Waze (running in the background to warn of speed traps) and Apple Maps (if I'm driving something with CarPlay and the infotainment's built-in navigation sucks).
But it left me wondering: how does it work? Paul Konieczny, CEO of Live Roads, gave me an explanation -- up to a point. "Primarily it is based around sensor fusion and certain probabilistic models -- we call it the Black Box," he said. "The current release of the app that is available in the Play Store has an earlier revision of our Black Box. This initial version is missing some of the functionality of the full-fledged system and thus has a spatial resolution of ~2.5m. This compares favorably to standard GPS that has a resolution of 4.0 m+." By summer, Konieczny hopes that the system will be fully operational and that accuracy will be down to under 1.5m. Assuming a large enough user base, that should let it offer lane-specific traffic data, "as well as introducing an entire ecosystem of 3D objects that users will be able to interact with," he told me.
But it left me wondering: how does it work? Paul Konieczny, CEO of Live Roads, gave me an explanation -- up to a point. "Primarily it is based around sensor fusion and certain probabilistic models -- we call it the Black Box," he said. "The current release of the app that is available in the Play Store has an earlier revision of our Black Box. This initial version is missing some of the functionality of the full-fledged system and thus has a spatial resolution of ~2.5m. This compares favorably to standard GPS that has a resolution of 4.0 m+." By summer, Konieczny hopes that the system will be fully operational and that accuracy will be down to under 1.5m. Assuming a large enough user base, that should let it offer lane-specific traffic data, "as well as introducing an entire ecosystem of 3D objects that users will be able to interact with," he told me.
Re: (Score:3)
Yes, but can we be sure of his velocity?
Another yutz who doesn't know what accuracy means (Score:2, Informative)
I think he meant precision.
Re: (Score:2, Informative)
I'm not really sure...I think it may be you who doesn't know what those terms mean. Accuracy is how close your measurement is to the correct value. Precision is how much variance you have between multiple measurements. This article talks about being able to identify correctly which lane you are actually in, which sounds like accuracy to me.
I assume in the process they are also increasing precision of the readings, too. They are probably cross referencing accelerometer and compass data with high resolution m
Re: (Score:2)
Because.... in this case, some one could provide their readout of location to 20 decimals. Incredibly precise!
And, be off in real location by 300', so, not very accurate.
Re: (Score:2)
Only works in prison cells because it's a cell phone and not a mobile phone.
Hopefully Virtual 3D Objects (Score:5, Funny)
... an entire ecosystem of 3D objects that users will be able to interact with ...
I already have an entire ecosystem of 3D objects I am trying not to interact with while I am driving.
Remember Garmin? (Score:1)
Garmin was using this exact 'technology' 18 years ago.
And hence forth has Apple and Google.
What a fucking shit for brains submission.
What happened to Slashdot.
Re: (Score:2, Informative)
The difference is that a Garmin has 12 receivers and a standard cell phone has 3. 3 receivers is the minimum that you need to find your position but it is only accurate to about 150 feet. What I believe this software package is doing is storing additional 3 GPS satellites frequencies in memory and switching between the two sets of 3 frequencies as fast as it can to get to 1.5 meters or around 5 to 7 feet.
With 12 receivers, you only need 12 because of the 24 GPS satellites, 12 of them are going to be on the
Re: (Score:2)
Most good mobile phones today support not only GPS but also GLONASS and Galileo (which is still not fully operational, but when you get a fix on one of them, it increases the precision).
I typically get a 3-4 meter resolution if the weather is good, which is better than the Garmin device I used to have in the car.
They call it "black box" (Score:4, Interesting)
You can get a better GPS fix by having a base station at a surveyed point, but this isn't what they're doing, because it would be expensive. There are also corrections you can do based upon atmospheric conditions, but they say that wasn't good enough. They pretty obviously don't want to say how they're doing it, which makes sense for trying to corner the market. But I wonder how novel their methods are, and if they can really stay ahead for long.
They say it's by "sensor fusion" and probabilistic models. The sensors are presumably the accelerometer and maybe the barometer in a cell phone. They could have an idea of the nominal error of a sensor based upon the model of phone. The probabilistic models, I can only guess, come from having many users running the app at once so they can try to reduce the error using many measurements.
Once they narrow the precision of the fix to the size of a lane of traffic, they could start identifying the lane by the behavior of clients in that lane -- if everyone turns off the freeway at this point in this lane, it's probably an exit lane, that sort of thing. And those rules could become much more complex. Like, use a machine learning process and train it on a bunch of well-defined traffic situations.
But, two questions: first, can they compete with companies like Garmin and Google who have tons of money and clients, plus smart people who could try to reverse-engineer these improvements and roll them out basically for free? It could easily be worth it to crush competition. And, if the success of the probabilistic model is based upon having many clients, do they get this level of improvement everywhere, or only in higher-traffic areas like cities?
Re:They call it "black box" (Score:4, Interesting)
But, two questions: first, can they compete with companies like Garmin and Google who have tons of money and clients, plus smart people who could try to reverse-engineer these improvements and roll them out basically for free?
Perhaps their probabilistic model suggests that one of the Big Dogs will buy them out at a ridiculous Silicon Valley rate and all the top folks will be flush with lots of cash-ola.
Why? Because by itself, there's really no way to adequately monetize this app, the amount of money they can make selling it through the app stores or directly to the phone companies just isn't all that much comparatively speaking.
No, they want Google to buy them.
Re: (Score:2)
Yeah, I think you're right. It seems like the best case for them is that they are doing something non-trivial enough that Google can't replicate it immediately and just buys them instead.
Re: (Score:1)
Re: (Score:3)
You can get a better GPS fix by having a base station at a surveyed point, but this isn't what they're doing, because it would be expensive.
Although that *is* how the original civilian GPS companies managed to get around the military-mandated fuzzing of gps data originally being provided to said companies...
Re: (Score:2)
Driving a car, to a location, 1.5m accuracy who cares. What I want to know is the basic route and where to park when I get there, what the location looks like, what the entry to the carpark looks like and the best route to a carpark, close to the final destination.
High resolution, sure if it is a big building and I want a walking route to the required location. Does that 1.5m accuracy including vertical accuracy because sometimes buildings come in storeys.
Re: (Score:1)
The point is (if they actually had the high accuracy) that you can now develop per-lane traffic and route prediction, including carpool lanes, express lanes, etc.
Re: (Score:2)
reverse-engineer these improvements and roll them out basically for free
Have you heard of patents?
Re: (Score:2)
Re: (Score:2)
without publishing their methods, they can't get a patent
Patents are published after application [wikipedia.org], not before. Just because they didn't tell this reviewer the technical details doesn't mean that aren't planning to apply for patents, if they haven't already.
Re: (Score:1)
But, two questions: first, can they compete with companies like Garmin and Google who have tons of money and clients, plus smart people who could try to reverse-engineer these improvements and roll them out basically for free?
Who cares? I have a Garmin, and Garmin has become grossly incompetent. It's a toss-up whether you'll even be able to get your GPS maps to update correctly. I'm never buying another Garmin again. Google discontinues apps at the drop of a hat (any hat) so they don't have the staying power to bring an app that complicated to fruition. They're wasting their time redesigning the image gallery app every year instead.
Re: (Score:2)
I think your comment about modeling the behavior is almost certainly true.
If i just turned left, and there's only one left turn lane, then they can look back a few seconds and adjust GPS to reflect that. Similarly if you are driving down a two-lane highway and they can use the accelerometer to detect lane changes, then they can calibrate the GPS to reflect the "probably-right' lane in each case.
Re: (Score:2)
I don't think they're aiming to compete with Garmin and Goo
no mystery (Score:2)
its simply GPS fused with other sensors and thats it...
GPS/GLONASS since Android 7.0 (i.e., Nougat), users now have access to raw GNSS measurements – opening the door to higher-accuracy
google maps even does this by asking you to accurately calibrate your compass
the problem is the antenna on most smart phones is terrible compared to those fitted in cars plus the ability for a car to accurately predict that your actually in a car compared to the smartphone which cant know this for sure...
Re: (Score:2)
plus the ability for a car to accurately predict that your actually in a car compared to the smartphone which cant know this for sure...
Let's see, traveling 100 kph on an express motorway, probably in a car. But we can't know this for sure, the phone could be in someone's pocket while they are holding on to the hood and screaming for the driver to stop.
Re: (Score:2)
Black box...kind of like Theranos (Score:2)
Call me skeptical.
Re: (Score:2)
Re: (Score:2)
"Machine learning" is the new alchemy. It must be working but no one "knows what it does".
Re: (Score:2)
There are a lot of issues with this. If you don't know why something works, it might work for completely the wrong reason. I heard about a group of researchers who tried to teach an AI to tell the different between wolves and dogs. It got very good, and they thought their AI had everything figured out. But, they late discovered that it wasn't even looking at the dog. It turned out that all the pictures they presented of wolves had snow in them. And so, instead of distinguishing wolf and not wolf, it was ac
GPS accuracy is only 10% of the battle (Score:5, Insightful)
90% is keeping maps up-to-date. What good is 1.5m accuracy, if you encounter a road construction detour just constructed yesterday? Google Navigation works so well not because of its accuracy, but because they put so much time, money, and effort into updating their maps constantly. Even then, it's not hard to find examples of locations where Google Navigation is not quite up-to-date. Good luck "Live Roads"!
Re: GPS accuracy is only 10% of the battle (Score:1)
Google uses the public to update their maps, via the report issue function (and previously Google map maker)
Re: (Score:2)
This is true, but Google Maps also puts a lot of their own money into it. They employ at least 10,000 people in their maps division alone!
http://www.businessinsider.com... [businessinsider.com]
Re: (Score:2)
Re: (Score:2)
1.5 meters does not tell you what road you are on, but what lane you are in
This is theoretically true, but this requires that the maps themselves also be accurate to within 1.5 meters. From my own experience using navigation apps, mostly Google, I can tell you that they are NOT that accurate.
"I'm not that familiar with Android" (Score:4, Insightful)
Is this class signaling? A disclaimer in case someone mistake you for Android trash and not the awesome Apple bro you actually are? Because I can't imagine anyone supposedly "from Ars Technica" finding themselves mystified by anything on a modern Android device. That just doesn't compute.
Re: (Score:2)
Alternatively, they could just not be familiar with Android. Maybe they use a flip phone? A laptop is enough portable computing for a lot of people.
Re: (Score:1)
Re: (Score:2)
Except that it's (currently) Android-only software that arrived on a S7
Re: (Score:2, Informative)
Well, I'm not overly familiar with iOS, so I'd add a disclaimer if I was reviewing an iOS app. I'd expect someone who's primarily an iPhone user to give a similar disclaimer when reviewing an Android app. Same for a Mac user reviewing Windows software and vice versa.
Re: (Score:1)
Is this class signaling? A disclaimer in case someone mistake you for Android trash and not the awesome Apple bro you actually are? Because I can't imagine anyone supposedly "from Ars Technica" finding themselves mystified by anything on a modern Android device. That just doesn't compute.
Had the same thought.
"I am of course not very familiar with McDonald's, so don't consider this review of the Trump Burger a "review" ..."
Re: (Score:2)
What, because if you carry around an Android device, you're intimately familiar with every nook and cranny of iOS?
It's completely reasonable to think that if someone doesn't use it every day, they may not be as familiar.
Only 1.5m? (Score:2)
My 2016 cellphone has that via differential GPS plus Inertial sensors, and is down to 1 foot accurate. I use it for mining spot location all the time.
Re: (Score:2)
I can step into the exact same snake hole in the middle of nowhere every time I have tried with my eyes almost entirely on the screen (taking eyes off the screen to navigate around obvious things like boulders, over ridges and such.)
It's a MIL-SPEC phone so I expect at least some minimum level of build quality and functionality and have received every bit of it.
What accuracy are we talking about (Score:1)
The accuracy commitments do not apply to GPS devices, but rather to the signals transmitted in space. For example, the government commits to broadcasting the GPS signal in space with a global average user range error (URE) of 7.8 m (25.6 ft.), with 95% probability. Actual performance exceeds the specification. On May 11, 2016, the global average URE was 0.715 m (2.3 ft.), 95% of the time. GPS Accuracy [gps.gov]
User accuracy depends on many factors in addition to range accuracy, so the result is GPS Accuracy Levels [nps.edu] it
Evil uses for the data (Score:2)
That data can be really useful to automatically find illegal lane changes (e.g. overtaking over barrier lines, pushing into queues over lines that indicate that you can't, turning from lanes that it is not allowed)
Why? (Score:3)
There's a lot of negative things to be said about Google Maps and Waze, but I can't say I've ever had a problem with accuracy that wasn't a half a second blip that was completely irrelevant to the current navigation: e.g. Maps has problems recognising I merged into a tunnel and then blips back onto another road for a second before correction.
Whoop-de-do.
Re: (Score:2)
Agreed - although they all seem to suffer on certain streets in London which have tall building. They also don't work in tunnels ;-)
As for accuracy though, I suspect it's a matter of using the GPS fix + as much else as you can. If driving, then you must be on one of the car lanes, so plot to that and then watch the accelerometer and compass for any change in direction not attributable to the road itself. Clever stuff, but not that clever.
I can see this being pretty helpful in GPS navigation though. Right no
Re: (Score:2)
Only as good as the most recent satelites. (Score:2)
Marketing BS ... (Score:2)
I call BS on the explanation ... those "probabilistic methods" are already implemented by most navigation programs, placing you not at the location of the GPS lock, but on the nearest street considering direction of travel. ... even street cutoffs typically do not need to be down to that scale ...
Also, who needs 1.5m accuracy for navigation? I've not had any destination that required me to be exact down to anything less than 5m
Now, if they can spit out the coordinates actually down to 1.5m repeatedly for th
Where Am I? (Score:2)
"I'm not that familiar with Android" (Score:2)
GPS + Inertial navigation + road matching + magic (Score:1)
Virtual differential GPS? (Score:1)
Some of the error in GPS measurements of closely located receivers is correlated. There is more information in the relative measured location of two units than in the absolute measured position of one. (Differential GPS is based on this: an accurate measure of a receiver's location is made by comparing its relative position to a known receiver location.)
So... rather than estimating the location of each user independently, maybe they fold in the relative position of each user and solve their positions jointl