I don't know this for a fact but it is possible that since you have selected walking and are travelling at 2 to 3 mph the app might be set up to do GPS sampling as opposed to running continously. This would make a big difference to battery usage. It might be handled a bit differently if you were travelling at 15 mph.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Do you really need your iPhone all the time?
- Thread starter houssein31
- Start date
- Sort by reaction score
Julien
macrumors G4
Here is a known distance 6 to 6.1 mile run I did (also recorded on my Garmin 620 and done many times) recored using my iPhones pedometer. It is off by over 35%.
Here are a couple of articles to help you understand how pedometers work and also to understand their limitations and accuracies.
http://www.mrsicklerphysicaleducation.com/6thGrade/6th_Grade_Page_files/How do Pedometers Work.pdf
http://walking.about.com/cs/measure/a/pedaccuracy04_2.htm
http://www.medicinenet.com/pedometers/article.htm
Maps data is sent through the internet but your point of location must be determined by a measurement and a calculation. In this case the measuring /calculation *tool is GPS data. I'm sorry you don't understand this and it is far to complex to explain in a soundbite.
Also to add. The aWatch must be paired with your iPhone (using the aWatch App) to get any internet data. It has no method of connecting to the internet on it's own.
EDIT: Here is a good basic Apple article about how location services work.
*There is a less accurate method using cell tower triangulation but the aWatch doesn't have a cell radio.
Last edited:
The inertial navigation systems on nuclear submarines can do this. But those things are huge (washing machine sized) and cost millions. Not something that can fit on a watch.
What the watch will give you is an inaccurate pedometer-like estimate. Maybe slightly more accurate if the watch includes a magnetic compass module.
I have a recent model Garmin GPS watch. It's big and clunky sized, and its tiny watch-sized rechargeable battery runs down in a few hours when the GPS is active. Not suitable for any slim watch specified for whole-day battery life.
You have.
iPhone 5S, for example. It only uses A-GPS (GPS+internet+cell) for initial fix, then it uses accels gyros.
Or how do you think indoor location works?
And yes, it's possible.
For example, missiles (including ICBM's) don't use GPS, too slow, to unprecise.
http://en.wikipedia.org/wiki/Inertial_navigation_system
Julien
macrumors G4
A-GPS is GPS plus cell phone triangulation and WiFi location data (if available) added together. GPS, Assisted by cell phone triangulation and WiFi location data or A-GPS
Rrriiiiight 😀
Satellite signals still get to you through windows. Get into basement and see if accels and gyros will navigate you there. 🙄
Unfortunately, iPhone is not a missile or ICBM.
Last edited:
So which App can trace a run without using GPS?
I have done some work previously on an iOS App, trying to calculate the speed of the device without using the GPS. You could maybe use it in a car on a flat road to tell when you have reached 60mph, but the accuracy was nowhere near good enough to trace a run around a park.
The issue is mentioned in the wiki article you linked to, small errors add up over time to mean massive errors in the results. Without some other sensor i.e. perhaps air speed in a missile I don't see how it would work.
You realise if you are sitting in a car at 100mph and sitting at your desk the accelerometer will give you the same reading? So if while you are in the car at 100mph you have calculated a speed of 105mph, you are now consistently 5mph out. And when you stop you may have a calculated speed of 5mph. Unless you correct it and tell the device every time you stop (also mentioned in that wiki article).
So yes you can use it for navigation ideally with other sensors, but can a watch with only an accelerometer and gyro trace a run around a park? Theoretically maybe but using the type of accelerometer in iOS devices I don't think so..
Last edited:
What's on the Y axis, and what does the varying color of the curve represent?
And double facepalm at
for the choice of grid/label intervals... 2.8... 2.9 miles? Srsly, ? You have monkeys there pounding on keyboards? Always wondered what crackhead coded the Stocks app axis, now I see the pattern. ****ing idiots.
Julien
macrumors G4
It would be speed. It was on an extreme trail run.
Don't understand what you rant is about. it is a Nike App and I don't know what algorithms Nike uses to pick the grid pattern or how/why it relates to Apple??? 😕And double facepalm at
Yes right, do you have a problem with that?
Yes, they can get through everything, except that the signal is so attenuated that a phone can't do anything with it.
But it doesn't seem you understand that.
And yes, you can get into basement, and see the tracking work.
Who said it was.
----------
Any App, it's given by the OS, the App is transparent to that.
That's because you aren't using the full spatial/temporal resolution an the full array of accelerometers.
iOS has access to that.
Not a problem for the application (recording workouts).
While on that, for human movement, civilian's GPS's have too low resolution (circa 1 meter, best), so accelerometers are in fact better.
Exactly, that's the problem of using inertial navigation.
That's why starting with the iPhone 5S, apple added the M7, to record the accelerometer at all the times. So it records the car acceleration to integrate the speed, and thus, your position.
Yes. it can. Doesn't mean it's a watch or a missile.
The problem is that you don't understand the difference from a simple pedometer to a COMPLETE INERTIAL NAVIGATION SYSTEM, complete with a dedicated processor for it, and software running all that.
Your Apple Watch-collected data will be uploaded to your iPhone, and your iphone will show you on a map where have you been running.
And that article from apple is outdated, meant for consumers, and even the developers.apple.com documentation is behind a wall, because those are trade secrets that even the developers don't need to know. Developers just set the desired precision and the OS works that.
That's not the point, location uses the internet for PTP/NTP to get a faster clock fix, if nearby Wifi routers/ibeacons are present, to get a estimate fix too, cell tower location data, etc.
You think you know everything and there's nothing more to learn, don't you? That's the problem.
Last edited:
No, but I'm currently learning the iOS SDK, because my master thesis (starts februrary 2015) will require me to write a mobile application, it's about connected tv and second-screen synchronization. I'm majoring in Telecommunications Engineering.
I understand how it works, and how sensors (including MEMS) work, how ADC's work, how digital signal processing works, etc.
I understand the theory behind it, I understand the limitations.
Good enough?
Maybe good enough for what you're doing, but not good for making up theories about how iPhone gets its location.
You seem to have no experience with integrating noisy signals and no understanding of basic physics. Since you're starting to learn iOS development, download the Accelerometer sample code from Apple and modify it to do the following: run accelerometer at 100Hz, for each axis (x, y, z) take current accelerometer reading (ax, ay, az) and calculate current xyz components of velocity and position, starting with zero velocity and x=y=z=0. You know, basic kinematics calculation. dVx = ax*dt, dX = Vx*dt + ax*dt^2/2. Output xyz positions on screen, put the phone on table and push the button. Watch the numbers. When you realize how wrong you are, come back and say so. 🙄
Think why iPhone never prompts you to calibrate accelerometer and gyro?
lol, you didn't expect did you?
"integrating noisy signals"
That basically sums up telecommunications. nice try.
As for basic physics, I took classical physics, electromagnetics, and waves, you're nobody to tell me I don't have basic knowledge of physics.
And no, I'm not using apple's accelerometer API, because that's not what the location API uses, the location API has direct access to hardware. The accelerometer (and specially gyroscope) are low-pass filtered for app's, because of security reasons (Google it up).
And why didn't you take the M7 into equation? Because it is a real time system just for sensors, wich enables a degree of temporal precision and accuracy simply NOT avaliable in a OS like iOS, specially on apple's app sandbox.
And for your little equation, it's not even complete, you've left out the data from the gyro, wich would help minimizing errors.
Proof that Location API internally uses accelerometer and gyroscope?
Proof? Have you actually seen accel/gyro data reported by iOS API? If not, run that Accelerometer sample and collect data in console using NSLog. Plot it and prove that a low pass filter has been applied to it.
Precision of what? Rotational matrix - yes. Translations, big movements - no. Do you have a proof that M7 can integrate your run around a park and not land you on the Moon after a half hour?
The equation is complete. This just shows your level of knowledge of physics. Gyro has nothing to do with the movement of the center of mass. Gyro is essential when determining orientation when the phone is rotated in place, but doesn't help determine its motion as a hole. Even orientation in space it gets wrong when the phone is moved with acceleration because it can only determine orientation relative to apparent gravity, but it doesn't know the direction of true gravity.
matrix07
macrumors G3
And double facepalm at
Do you know that it's NIKE+ app and the design has nothing to do with Apple's at all, right?
You do not know what I was doing do you? How could you?
Full array of accelerometers? Last I checked the iPhone has one 3-axis accelerometer.
Again my posts are based on my own experiments. I have apps using the location and motion API.
Perhaps you should do as samiznaetekto suggests and create a sample App and actually experiment with this stuff..
Create an App and try to trace a route around your room, let alone a park.
Last edited:
Julien
macrumors G4
Science Fiction: [aWatch having some mythical] "COMPLETE INERTIAL NAVIGATION SYSTEM" without GPS and Unicorns
Science: Recording a track requires accurate longitude, latitude and (optionally) altitude recorded at precise intervals (usually 1 second). The aWatch requires paring with the iPhone’s GPS radio in order to recored this. The aWatch’s internal accelerometer and gyroscope are capable of only estimating distance and speed with NO tracking info recorded. Even Apple confirms this.
Apple Watch Press Release
Apple said:
Last edited:
IGNORANCE: this post.
Apple confirms it uses sensors in conjugation with GPS from the phone. That's a fact.
This is science, people who "believe" like you, aren't needed, except for your $$$ if you want to buy anything you don't understand how it works.
----------
Simple, use the iPhone inside a faraday cage. Enough proof.
Proof that it's not.
And there's proof that a low pass filter has been applied to it, because in previous iOS versions, it wasn't.
Civilian's gps is only good for about 1.1 meter precision, the iPhone can track better than this.
no, it's not.
You have zero knowledge of physics and real world limitations.
----------
Last time I checked, the world is 3 axis, so the iPhone has what's needed.
My posts are based on scienctific knowledge, what I know about apple, and my own experiments.
I suggest you to acces the raw sensor data, the same the os has acces to.
So why did you state there is an array of accelerometers?
I can only suggest you go create an App that does what you are suggesting. If you can't either because you do not know how, or because it cannot be done you really have no place slamming everyone who disagrees with you. If you can just post it up here thanks.
Julien
macrumors G4
Occam’s razor:
1 aWatch is using an unknown system to record coordinates that only you know about but can’t explain (see A).
2 The aWatch uses the iPhone’s GPS to track just the way Apple (see B), all tech pundits, all tech articles, all fitness articles and every one on this forum says.
A: You leave you iPhone at home to go for a run. You drive to latitude/longitude 38.749999 -9.125644 (perhaps familiar) to start your run. How does the aWatch acquire the staring latitude and longitude to track you?
B: Apple Watch Film-Health and Fitness
Time Code 34 to 50 seconds.Jay Blahnik Director of Fitness said:
Last edited:
Julien
macrumors G4
Excellent explanation and extremely informative video.😎 Thanks for posting. I'm sure Tanegashima will say it's "IGNORANCE".😀
EDIT: Couldn't stop watching and went to the end.😉
Last edited:
flur
macrumors 68020
Well, I'm not an engineer or an iOS developer, but I own an UP band and a TomTom GPS watch and have run with both of them multiple times. With a programmed stride length, the UP band, with just it's accelerometer, is darn accurate to the distance my TomTom reports. So I'm calling BS on the claim that accelerometer distance measurements are poor and always off. When done right, they can be quite good.