Most accurate GPS Tracking App

Forums › Android › Most accurate GPS Tracking App

curlynz

16 posts

Geek

#75265 15-Jan-2011 12:37

Hi all, I am looking for a GPS tracking app that has the best accuracy with following twisty routes. I have been trying a few and am finding that the accuracy varies greatly. Am finding that on twisty roads or tracks they will end up plotting a straight track instead of actually following the curve

Am using a Samsung Galaxy S running 2.2 so know the accuracy with the phone is not the greatest to start with

Cheers

1 | 2

4431 posts

Uber Geek
Inactive user

#427395 15-Jan-2011 12:59

Have you tried "Maverick"? http://www.appbrain.com/app/maverick/com.codesector.maverick.lite

There's a pro version too

n4

959 posts

Ultimate Geek

#427408 15-Jan-2011 14:17

I've had pretty good results with Endomondo on the SGS. Make sure the SGS setting for 'Use Sensor Aiding' (Settings->Location and Security) is off, I think that tries to smooth out the SGS GPS performance and can cut corners/overshoot. But then you will have to have an uptodate ROM with improved GPS performance.

Samsung Note20 Ultra, on 2degrees

nitrotech

1285 posts

Uber Geek

#427424 15-Jan-2011 15:13

I find runkeeper pretty good - pro version is free at the moment - grab it quick!

oxnsox

1923 posts

Uber Geek

#427472 15-Jan-2011 17:57

What accuracy do you expect/need???

muppet

2568 posts

Uber Geek

Trusted

#427486 15-Jan-2011 18:31

Accuracy has little to do with the app and lots to do with how sensitive the GPS receiver is. That said, how often it plots where you are will also affect things.

The GPS receiver in a lot of mobile phones isn't that sensitive (a sensitive chipset is more expensive)

I have the same problem with my dedicated Garmin Handheld GPS. If I jog under shop awnings then the GPS "guesses" where I am. It's because it doesn't have much signal under an awning. Winding Roads in a valley are notorious for causing this sort of problem. There's signal echo and all sorts of issues to deal with.

Audiophiles are such twits! They buy such pointless stuff: Gold plated cables, $2000 power cords. Idiots.

OOOHHHH HYPERFIBRE!

Linuxluver

5828 posts

Uber Geek

Trusted

Subscriber

#427510 15-Jan-2011 20:53

curlynz: Hi all, I am looking for a GPS tracking app that has the best accuracy with following twisty routes. I have been trying a few and am finding that the accuracy varies greatly. Am finding that on twisty roads or tracks they will end up plotting a straight track instead of actually following the curve

Am using a Samsung Galaxy S running 2.2 so know the accuracy with the phone is not the greatest to start with

Cheers

I have a Galaxy S and a Nexus One and a Vodafone 845.

I have used Google Maps / Nav on all three.

The GPS functionality on the Nexus One leaves the Galaxy S for stone cold DEAD.

I was in Canada over the past month. I didn't rent a car and instead took trains everywhere. I compared the Nexus One, the (cheap, lowly) Vodafone 845 and the Galaxy S over and over. There is NO DOUBT in my mind, having seen it, that the Galaxy S is the worst of the lot. Even the Vodafone 845 peformed well compared to the Galaxy S.

But King of the GPS was the Nexus One. I would be sitting on a moving train doing 100km/h or more...and the Nexus One Maps / GPS would be tracking me - literally - ON the train tracks...and when we crossed a road or river...the map was 100% in sync, showing me crossing the river in the same second....consistently locating me within as little as 5 metres despite my speed and the fact I was sitting in a (nice, comfortable) arm chair inside a train carriage.

Provided the Nexus One can see a GPS satellite, you will know *exactly* where you are. The Vodafone 845 was almost as good.

Good luck with the Galaxy S. For GPS, it's actually poor. It works.....but it's not in the same league as the Nexus One or the lowly Vodafone 845. The Galaxy S is *awesome* in many respects and I love mine to bits.....but GPS isn't one of the reasons. Driving home this afternoon, the Nav would be fine....then suddenly I'm a km away from where I actually am and the Nav re-routes....(WTF!?) then a few seconds more and it's realised I'm where I am and re-routes again....correctly this time. It happened twice on the way to the place I went to in Titirangi and three times on the way back. It's only a few seconds each time.....but the randomly wrong Nav instructions issued every few minutes shake one's faith.

I've never had that experience with my Nexus One.

_____________________________________________________________________

I've been on Geekzone over 16 years..... Time flies....

muppet

2568 posts

Uber Geek

Trusted

#427511 15-Jan-2011 21:05

Linuxluver: Provided the Nexus One can see a GPS satellite, you will know *exactly* where you are.

GPS needs a minimum of 3 satellites to give a fix on where you are (triangulation)

But that's probably being a bit too pedantic for what you meant :)

Audiophiles are such twits! They buy such pointless stuff: Gold plated cables, $2000 power cords. Idiots.

OOOHHHH HYPERFIBRE!

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

oxnsox

1923 posts

Uber Geek

#427516 15-Jan-2011 21:28

GPS, by design, is really only good to about 15m but thats good enough to be exact in most instances because it put us in eyeball range, which gives the appearance of exactitude.

Beyond that, and here LL can probably add comment, there's a lot that can be done with software to 'smooth' or 'manipulate' your position, so would be interested to know if you were using the same map-app on your 3 devices.
[Smoothing can use info from other fone sensors along with a smarter algorithm so that your track isn't a series of straight lines between sat-fix positions whilst other app based 'manipulations' will lock your car to the correct side of the road (or rail-track) for your direction of travel.]

But as said the key determinant is signal quality, and that's pretty much determined by the GPS antenna in the device (and how much 'noise' the other handset functions create). The actual receiver (chip-set) qualities are all pretty comparable at this end of the market.

Linuxluver

5828 posts

Uber Geek

Trusted

Subscriber

#427532 15-Jan-2011 23:40

oxnsox: GPS, by design, is really only good to about 15m but thats good enough to be exact in most instances because it put us in eyeball range, which gives the appearance of exactitude.

Beyond that, and here LL can probably add comment, there's a lot that can be done with software to 'smooth' or 'manipulate' your position, so would be interested to know if you were using the same map-app on your 3 devices.
[Smoothing can use info from other fone sensors along with a smarter algorithm so that your track isn't a series of straight lines between sat-fix positions whilst other app based 'manipulations' will lock your car to the correct side of the road (or rail-track) for your direction of travel.]

But as said the key determinant is signal quality, and that's pretty much determined by the GPS antenna in the device (and how much 'noise' the other handset functions create). The actual receiver (chip-set) qualities are all pretty comparable at this end of the market.

My Nexus One can put me in my chair, on my deck....maybe leaning one way or the other...definitely within 2 metres four times out of five....if not better

The Galaxy S is usually out by about 15 metres and typically errs to the North....putting me somewhere in the house of my northern neighbour instead of on my front deck. This is fairly consistent....except when it puts me in the house of the neighbour to the south....or a kilometer or more away in any direction....apparently randomly. To be fair, it works well enough to be useful and I have used it. But when you've seen the Nexus One (or the 845) do so much better....you realise how good the Galaxy S *SHOULD* be for a $900-$1000 phone.....and isn't.

_____________________________________________________________________

I've been on Geekzone over 16 years..... Time flies....

Linuxluver

5828 posts

Uber Geek

Trusted

Subscriber

#427539 16-Jan-2011 00:06

I just did this test. Standing in the same place, which phone most accurately located me? The Nexus One or the Samsung Galaxy S? I took screen shots - both taken after several minutes to allow the location to settle down and locate me as accurately as possible. I ignored "rogue" results as much as possible. I had to photograph the Nexus One as it isn't rooted and I can't take screen shots on it.

Once it settled down, the Nexus One consistently located me right where I was standing: just outside my front door. Bang on. Nailed it. Within a metre. Even while it was settling down, it never had me more than 10 metres from where I was standing and gradually became more precise over a couple of minutes.
Nexus One showing my Location in Mapsocation

Nexus One showing my Location in Mapsocation

Meanwhile....the Samsung Galaxy S used every one of those 30 metres....placing me variously two houses to the North, across the street inside that neighbour's house, up the road...and more or less anywhere in any direction 30 - 50 meters from my front door. This screen capture taken after several minutes shows me standing on the rear deck of my rear neighbour's house...a good 40-50 metres away from where I was standing. It's representative of the level of (in)accuracy of the Galaxy S GPS. Even 15 minutes later it rarely got as close as 10 metres...and usually more like 20 or 30...as the bubble says.

This test isn't exhaustive or definitive.....but it is a good indication of what I have been seeing for several weeks.

Samsung Galaxy S is much less accurate.

_____________________________________________________________________

I've been on Geekzone over 16 years..... Time flies....

oxnsox

1923 posts

Uber Geek

#427553 16-Jan-2011 07:55

Thanks for that LL, all actually very interesting.

From that it appears the Galaxy is using pretty raw fix data with not a lot of averaging or smoothing, whilst the Nexus has got some extra good GPS algorithms built in. (its implying an accuracy that isn't available without augmentation of the fix)

I wonder how much of that may be due to Google and the flagship status of the Nexus product.... would be interesting to see how the Galaxy Nexus performs.

muppet

2568 posts

Uber Geek

Trusted

#427561 16-Jan-2011 08:41

oxnsox: Thanks for that LL, all actually very interesting.

From that it appears the Galaxy is using pretty raw fix data with not a lot of averaging or smoothing, whilst the Nexus has got some extra good GPS algorithms built in. (its implying an accuracy that isn't available without augmentation of the fix)

How does that work? Wouldn't smoothing/averaging of data remove accuracy, not increase it?
Is your statement based on anything you've read, or just how you feel GPS works?

More likely is that the Samsung Galaxy S is able to listen to less channels at once (http://www.gpsreview.net/gps-chipset-channels/) or it has a less sensitive reciever/chipset, or is less able to deal with Multipath errors and is therefore unable to figure out with as much accuracy.

Is either phone able to give you a display of the number of satellites it can see? For example this is what my handheld GPS shows me.

Audiophiles are such twits! They buy such pointless stuff: Gold plated cables, $2000 power cords. Idiots.

OOOHHHH HYPERFIBRE!

fahrenheit

757 posts

Ultimate Geek

#427634 16-Jan-2011 13:50

Despite all the software fixes out there for the Galaxy S, mine is still underperforming (rooted with custom rom/fixes).

I think that the antenna fix, which involves bending up a little piece of folded metal is probably the only option. Samsung did change this component (post-Oct builds if I recall correctly). With the different hardware revisions out there, its difficult to know whether reports of good or poor gps performance are based on this change.

There is a thread on xda which covers this mod in detail (not that I can seem to find it at the moment) and I'd like to give it a go when I have time.

oxnsox

1923 posts

Uber Geek

#427709 16-Jan-2011 17:59

muppet: Wouldn't smoothing/averaging of data remove accuracy, not increase it?

As you know GPS positions are derived from very accurate clock data sent from the satellites, with the actual GPS chipset using 'time-of-flight' calculations to measure it's distance from the satellite (remember time and distance is always changing). Whilst the concept is simple it is a complex process as a number of very small variations affect both the accuracy of the data and, more obviously, the path it takes. (eg: the signal does note follow a direct shortest-line path due to refraction thru the earths ionosphere/atmosphere and thence trees/buildings etc.)
With raw GPS data (and most early units) you can see this as no 2 subsequent position fixes will be in the same spot and you'll appear to jump all over the place. Using simple averaging processes manufacturers can negate much of this wandering and give the appearance of better accuracy. When you move with a GPS unit the direction of travel also has an averaging effect on transverse errors.
With better processing power, smarter algorithms, and integrating and referencing associated motion sensors you'll also get some improvements.

Is your statement based on anything you've read, or just how you feel GPS works?

I've been working with GPS for over 20years

More likely is that the Samsung Galaxy S is able to listen to less channels at once (http://www.gpsreview.net/gps-chipset-channels/) or it has a less sensitive reciever/chipset, or is less able to deal with Multipath errors and is therefore unable to figure out with as much accuracy.

Channel talk is probably more marketing hype than anything much these days. In the early days of GPS when the going rate was tens of thousands of dollars channel numbers mattered (even though there were less satellites about than receivers had channels for). Garmin pretty much re-wrote the book by introducing a unit that could sequence, very fast, thru multiple channels and provide similar performance at much lower cost.

Most import factor to getting good usable signals is antenna design, years ago the DSIR (or one of its off-shoots) did a lot of research on this with the most obvious benefit at the time going to then fledgling company Navman. For really high accuracy solutions that minimise multipath issues the aerials aren't something you'd be carrying round with you.

curlynz

16 posts

Geek

#427996 17-Jan-2011 13:22

Have been using Endomondo and Runkeeper and have found that Endomondo seemed the more accurate of the two. Even tried them with Trackerbooster running in the background with is meant to help.

Decided to upgrade to 2.2.1 through Kies which was super easy with the Kies registry Patcher software and then decided to go one step further and have chucked on Darkys Rom V8.1.

Have noticed that my actual location seems to be more accurate, what I am actually looking for is tracking software that plots points at a greater rate hence picking up turns etc. I have noticed in endomondo when you can watch it you can see that the gps is updating often as you go around a bend but the algorithm to plot the line will then cut out points thus making it straighter and actually cut out the corner.

Been o\investigating more and have d/l Maverick, Sports tracker and My Tracks. The last two have options to lower min point interval time and min point distances Might be what I am looking for. Will have a play and let you know how I go.

Oh yeah found something else that maybe out of interest to some especially that have pre Oct 10 Samsung Galaxy S phones. Its a little hardware mod for the gps antenna for those that are brave.

1 | 2