[aprssig] iPhone web position

Steve Dimse steve at dimse.com
Sun Aug 2 21:06:31 EDT 2009


The cell position is usually not the exact location of a tower. Towers  
have directional antennas, so there is usually at least a 120 degree  
sector to narrow it, they use signal strength and often you are heard  
by more than one tower. I've seen some very weird guesses though that  
could not be attributed to any pattern heard by one or two towers, so  
it is far from a perfect approximation.

There is a way to get better fixes, and I'm working on it. Essentially  
you tell the phone to send positions continuously. It does exactly  
that, if I just replace the function call the positions come so fast  
neither the radar map nor the position update cgi get called. That is  
why I have to finally break down and learn javascript, to filter out  
the positions and only update when there is a significant change. I  
will make it work though, promise. It will need to be an option, if  
you've ever used something that runs the GPS continuously it drains  
the battery fast. It'd be great for running on power in a car though.


Native apps have a very sophisticated API to allow you to set minimum  
accuracy, and to send a new posit only when it changes n meters. That  
is what my ill-fated app and ibcnu use. That is not available in the  
web api.

Steve K4HG

On Aug 2, 2009, at 8:52 PM, Wes Johnston, AI4PX wrote:

> This works for me!  8gig iphone 3g.  I do wish it had a way to make  
> the GPS stay on longer before reporting. Google Latitude has the  
> same issue... reports posit too soon and the position is a very  
> rough guess.  The funny thing in my case is that the postion  
> reported is NOT my local cell tower, I'm actually about 1.5 km from  
> my tower and it's reported me on the other side of the tower by  
> 1.5km.  The IBCNU app does have a switch to flip to limit the  
> minimum accuracy.  Gues I'll have to do some reading on the  
> geolocation api.  I did see something in there about age of a cached  
> posit but nothing about accuracy that I saw in my cursory reading.
>





More information about the aprssig mailing list