Loading ...
Sorry, an error occurred while loading the content.

Re: [aprsisce] Position Ambiguity

Expand Messages
  • Lynn W Deffenbaugh (Mr)
    ... I do have it on the ToDo list. I asked the question so I could understand the use of such a function as well as to get an idea of what priority to assign
    Message 1 of 20 , Jan 1, 2011
    • 0 Attachment
      James Ewen wrote:
      > On Sat, Jan 1, 2011 at 9:19 AM, Lynn W Deffenbaugh (Mr) <kj4erj@...> wrote:
      >
      >
      >> Not yet. I'm still not totally convinced as to the utility or
      >> usefulness of ambiguity and it seems to always raise the question of
      >> "it's not showing me where I really am". I'll add it to the list to
      >> support in some future revision.
      >>
      >
      > This to me is a strange mindset. You are basically saying I'm only
      > going to implement a subset of the APRS specification, the parts that
      > make sense to me.
      >

      I do have it on the ToDo list. I asked the question so I could
      understand the use of such a function as well as to get an idea of what
      priority to assign to it. I plan to support all of aprs101.pdf and as
      much of the 1.1 and 1.2 addendums as I can find specific documentation
      (and usefulness) for.

      > Obviously there's no way to be able to make a program support every
      > single aspect of APRS from the start unless you sit on it under
      > development until it is perfect, then release it. Obviously we're
      > playing with the program as it is growing, and we get access to new
      > features as they are added. We suggest features desired, and wait for
      > them to bubble to the top of the queue.
      >
      > Position ambiguity would seem to be a fairly simple feature to
      > implement. I'd hang it in the menu structure in the Precision section
      > where you have the !DAO! feature. Of course that now brings up the
      > query... if you don't see the usefulness of ambiguity, why have the
      > ability to disable 2 digits of !DAO! precision? If you can resolve a
      > position accurately to give the extra precision that !DAO! extensions
      > give, then why have the ability to shut it off?
      >

      Every feature is "easy" to implement. But if you've got a thousand
      small rocks to move, it still requires a substantial amount of time to
      move them all.

      > Obviously I'm playing the Devil's Advocate here, just trying to show
      > the other side of the coin.
      >

      I appreciate the conversation, truly.

      > One thing that I love is that you draw the purple circle around the
      > received stations that have ambiguous positions. I don't know of
      > anyone else highlighting ambiguous position reports. That's the
      > biggest problem with position ambiguity... it's not that people use
      > it, but rather that receiving stations ignore position ambiguity, and
      > add digits of precision which are inaccurate, and then depict the
      > inaccurate location to the end user as if it were a high precision
      > location report.
      >

      The purple circle was done in self-defense a while back when I received
      complaints about a station that "wasn't showing the proper location on
      the map". It took a few days of research before I finally noticed the
      ambiguous position that the station was transmitting. So, now there's a
      way to highlight it on the map and it shows in the station popup dialog.

      > I sent a feature request in to Hessu at aprs.fi yesterday suggesting a
      > method of indicating position ambiguity on the site. I like your tight
      > purple circle idea, which highlights the station as transmitting an
      > ambiguous position. I'd put the station at midrange values rather than
      > padding out with zeros so as to minimize the amount of possible
      > maximum displacement. Then I would highlight the rectangle of possible
      > locations where the station would be, rather than draw a circle.
      >
      > When zero padding, the icon ends up on one vertex of the rectangle of
      > possible locations. Drawing a circle at that location then puts 3/4 of
      > the circle outside of the area where the station could possibly be,
      > which again is adding to the depicted inaccuracy.
      >

      My circle is actually centered by filling in 1/2 of the missing digit.
      If it is bb.bbb, I draw it at 30.00. If it is xb.bbb, it ends up at
      x5.000. xx.bbb is xx.500 and so forth.

      > Having the ambiguous station icon drawn at the center of the ambiguity
      > rectangle with a tight purple circle around it, and then when zoomed
      > out far enough being surrounded by a rectangle showing the bounds of
      > the ambiguity area would be the 100% best solution in my mind. This
      > would alert the observer to the fact that the station is somewhere
      > within the bounds of that rectangle.
      >

      I considered a rectangle, but the circle looked cooler. And for the
      purists, they can just square up the area with their eye.
      > Obviously if one were to select position ambiguity in APRSISCS/32, the
      > program should draw the area of ambiguity around the user's location
      > so the user realizes that they are transmitting lower precision
      > location information. That would actually look interesting while
      > driving around... rectangles would highlight as you drive from one to
      > the next.
      >

      Interesting idea, but I'm still probably going to insist on Me being at
      the best known actual coordinate. However, there's another ToDo item
      that will show the last point at which a beacon was transmitted, and
      possibly also show the last point where a digipeat was heard. The last
      transmission point would be a good place to show the ambiguity. Even
      for a non-moving station, it could then be offset by up to 1/2 the
      ambiguity (or something like that...).

      > Just my 2 bits...
      >

      As long as you're not looking for change! <grin>

      Lynn (D) - KJ4ERJ - Author of APRSISCE for Windows Mobile and Win32

      > James
      > VE6SRV
      >
    • h0chus_p0ckus
      Maybe we could come up with a different visual representation for stations with ambiguity set? Circle of ambiguity is I believe the phrase used. The
      Message 2 of 20 , Jan 1, 2011
      • 0 Attachment
        "Maybe we could come up with a different visual representation for
        stations with ambiguity set?"

        "Circle of ambiguity" is I believe the phrase used. The conventional graphic representation when you are unsure of the position, is to use a shaded circle which is solid at the center (the reported or probable position) and fades out toward the edges of the probable position. You'll see this used in cell phone mapping applications, where the position may be 'exact' from a GPS, or somewhat less exact from a SkyHook assist, or really uncertain, from a cell phone tower triangulation. The "certain" position is a spot, the less exact one a shaded circle, the least certain one a much larger shaded circle.

        The answer is already out there, the rest is just a simple exercise in programming. (Sorry, Lynn!<G>)

        --Red
      • ka0azs
        One reason is to show that there is a digi in that area. Helps with event planning (do I need to dig up a portable digi and all the work set up for a temporary
        Message 3 of 20 , Jan 1, 2011
        • 0 Attachment
          One reason is to show that there is a digi in that area.

          Helps with event planning (do I need to dig up a portable digi and all the work set up for a temporary set up entails?)when you know there is coverage in your area of interest. Particularly when setting up support in area you don't frequent on a daily basis.

          And, quite frankly, I didn't even consider the possibility of NOT putting a position in. Never occurred to me.

          I'd have to go back and look at my notes (if I can find them, been almost 10 years), but we were doing this with a burned EPROM, and I'm not sure if the configuration file let you do "no position", or if it would have simply broadcast a position of 0.00 by 0.00.

          73

          Randy KA0AZS
          --- In aprsisce@yahoogroups.com, Rob Giuliano <kb8rco@...> wrote:
          >
          > Why do you need to have the digi provide its position at all?
          > I realize that information is a valuable part of the system, but if it means no digi vs digi, is it a must?
          >  
          >  
          > It has to ID, but I don't think it "must provide" its position.  At APRS.fi and such, it might be hard to follow, but the diretive of getting the data  across the "dead spot" would still be accomplished.
          >
          > Robert Giuliano
          > KB8RCO
          >
        • ka0azs
          ... I would find it useful at public service events where I m setting up a portable station with no GPS and limited time frame to fool around with it. Of
          Message 4 of 20 , Jan 1, 2011
          • 0 Attachment
            --- In aprsisce@yahoogroups.com, "Lynn W Deffenbaugh (Mr)" <kj4erj@...> wrote:
            >

            >
            > Ok, finally a justification that I can understand and appreciate.
            > Question is, how many people will really set and unset Ambiguity >for that purpose? And with APRSISCE (the most popular mobile use), >they're probably running a real GPS and direct APRS-IS connection, >so they're probably not approximating their position anyway.

            I would find it useful at public service events where I'm setting up a portable station with no GPS and limited time frame to fool around with it. Of course, the APRSIS32 version of using the map and cross-hairs to set a position might make that obsolete.

            I'll have to play with it over the next year as I support PSEs. In the past the position was entered manually with coords from the nearest available GPS (sometimes passed by voice from a nearby vehicle) or pulled from a map. I'm not sure about +SA, but IIRC DOS automatically kicked in and calculated the ambiguity based on number of significant digits in the coordinates.

            As I've stated before, I'm very RF centric. Don't have any phone that will support the CE version, no plans to get one in the near future, refuse to pay for dataplans (I only have a cell because work requires it and family insists). And, as I consider PSEs training for emergency operations, I always assume the internet will NOT be available when I need it, hence I play with the IS parts only as an afterthought.

            > It was somehow a disguised tower? This is exactly the kind of >thinking that makes absolutely no sense to me.

            Wasn't a radio tower. A very tall building with some sensitive types working in there (as in their activities were considered sensitive, not that they got hurt feelings easily). I think it was also a bit of "we're willing to do it for you, but we don't want everyone up here asking for space".

            > I haven't done any foxhunting or DFing, but I thought the general >idea was to start away from the target and triangulate a bearing >from multiple positions decidedly distant from the target?

            Not an expert, but by my understanding, DFing the signal (which would be fun, isolating that one transmitter on 144.39) from a distance would resolve it into a position fix approximately the same size of the ambiguity circle. Moving into that area (not all of which is accessible to the public) to further localize it would attract the attention of several folks, the nicest of which would be our friend in Dispatch. Probably over thinking it, most folks would look at the transmitted position and call it good. but it made the site owners happy.

            >...and such a site was probably a TNC beacon a static beacon?

            Yep. AEA with the config file burned into an EPROM

            > Just guessing and trying to learn, not challenging or being >obstinate (although it sure reads that way to me).

            Not to me. If nothing else these type of questions might help me handle a similar situation in the future with a better approach

            >Questions stimulate discussion and with discussion comes learning >and understanding.

            Absolutely

            73

            Randy KA0AZS
          • Lynn W Deffenbaugh (Mr)
            WinLink s grid-square system doesn t show up in APRS. Something somewhere WINLINK is injecting 0.1nm abiguity position packets into APRS-IS for the
            Message 5 of 20 , Jan 1, 2011
            • 0 Attachment
              WinLink's "grid-square system" doesn't show up in APRS. Something
              somewhere "WINLINK" is injecting 0.1nm abiguity position packets into
              APRS-IS for the WinLink stations These packets specify a lat/lon, not a
              gridsquare, so the distinction is lost in the translation. (See
              http://aprs.fi/?c=raw&call=WINLINK&limit=50&view=normal) I can only
              understand and appreciate things that I can see in the lens through
              which I view the world (APRS-IS).

              Lynn (D) - KJ4ERJ - Author of APRSISCE for Windows Mobile and Win32

              Jordan Hayes KG6UAE wrote:
              >> Ok, finally a justification that I can understand and appreciate.
              >>
              >
              > So you don't understand and appreciate the WinLink idea of publishing
              > grid-square for a system, where the area that it serves is more
              > important than exactly where it's located?
              >
              > /jordan KG6UAE
              >
            • Jordan Hayes KG6UAE
              ... I think you re being obtuse. ... I think the lat/lon is the center-point of the grid-square in this case. Since APRS doesn t know from grid-square, that s
              Message 6 of 20 , Jan 1, 2011
              • 0 Attachment
                Lynn writes:

                > WinLink's "grid-square system" doesn't show up in APRS.

                I think you're being obtuse.

                > Something somewhere "WINLINK" is injecting 0.1nm abiguity position
                > packets into APRS-IS for the WinLink stations These packets specify
                > a lat/lon, not a gridsquare, so the distinction is lost in the
                > translation.

                I think the lat/lon is the center-point of the grid-square in this case.
                Since APRS doesn't know from grid-square, that's the best the WinLink
                folks could do. My point is that the lat/lon of those stations should
                probably not be represented on the screen in the same way as a tracker
                with 6-digits of accuracy are represented.

                > I can only understand and appreciate things that I can see
                > in the lens through which I view the world (APRS-IS).

                Yes, this is exactly what I am asking you to do: see the world through
                the lens of APRS and see that a different representation of ambiguous
                positions be used in APRS.

                You said earlier that you don't understand or appreciate it. You're not
                trying hard enough.

                Sheesh,

                /jordan KG6UAE
              • James Ewen
                ... More likely circle of probability is more likely. ... Yes, that would be a way to represent a possible location. However in APRS terms, the ambiguity is
                Message 7 of 20 , Jan 1, 2011
                • 0 Attachment
                  On Sat, Jan 1, 2011 at 12:53 PM, h0chus_p0ckus <reply.via.list@...> wrote:

                  >  "Circle of ambiguity" is I believe the phrase used.

                  More likely circle of probability is more likely.

                  > The conventional graphic representation when you are unsure
                  > of the position, is to use a shaded circle which is solid at the center
                  > (the reported or probable position) and fades out toward the edges of
                  > the probable position.

                  Yes, that would be a way to represent a possible location. However in
                  APRS terms, the ambiguity is defined by dropping digits of precision,
                  and because the location is described by using lat/long pairs, as you
                  drop digits of precision, you create rectangles of probability, with
                  the probability of being at any location within that rectangle being
                  uniform throughout the rectangle.

                  If you were to use shading to represent the possibility of the station
                  being at any one point in the rectangle, where the opacity of the
                  shading represents the possibility of the station being at that
                  location, you would need to use uniform opacity throughout the
                  rectangle.

                  Using a circle misrepresents the area of possibility, and using a
                  fading pattern within the circle could be interpreted to represent
                  that the location suggested is somehow of higher probability of being
                  the actual location than other areas within the shaded area.

                  Having any type of indication of stations sending ambiguous reports is
                  better than nothing, but if you are going to try representing
                  something, it is better to do it right than to go through multiple
                  iterations of incorrect representations, and having to explain to
                  people why:

                  a) the circle around the icon that is plotted doesn't include the
                  actual location they are at, or
                  b) the shading where they actually are located is lighter than near
                  the icon, when they know there's better probability of finding them
                  further away from the icon location.

                  I'm all about accuracy, even when it includes imprecise information.
                  Accuracy in my world is far more important than precision.

                  James
                  VE6SRV
                • James Ewen
                  ... That is incorrect... there are a number of cases where APRS packets can include Maidenhead grid square information. James VE6SRV
                  Message 8 of 20 , Jan 1, 2011
                  • 0 Attachment
                    On Sat, Jan 1, 2011 at 2:30 PM, Jordan Hayes KG6UAE <kg6uae@...> wrote:

                    > Since APRS doesn't know from grid-square,

                    That is incorrect... there are a number of cases where APRS packets
                    can include Maidenhead grid square information.

                    James
                    VE6SRV
                  • Lynn W Deffenbaugh (Mr)
                    ... I believe he was referring to my comment that the GridSquares known to WinLink were not being used in the APRS-IS packets transmitting the ambiguous
                    Message 9 of 20 , Jan 1, 2011
                    • 0 Attachment
                      James Ewen wrote:
                      > On Sat, Jan 1, 2011 at 2:30 PM, Jordan Hayes KG6UAE <kg6uae@...> wrote:
                      >
                      >
                      >> Since APRS doesn't know from grid-square,
                      >>
                      >
                      > That is incorrect... there are a number of cases where APRS packets
                      > can include Maidenhead grid square information.
                      >

                      I believe he was referring to my comment that the GridSquares known to
                      WinLink were not being used in the APRS-IS packets transmitting the
                      ambiguous location of the same station.

                      And, IIRC, there's only one non-deprecated use of GridSquares as a
                      location in APRS. Correction 2. From page 25 (35) of aprs101.pdf.

                      An alternative method of expressing a station’s location is to provide a
                      Maidenhead locator (grid square). There are four ways of doing this:

                      • In a Status Report — e.g. IO91SX/- (/- represents the symbol for a
                      “house”).
                      • In Mic-E Status Text — e.g. IO91SX/G (/G indicates a “grid square”).
                      • In the Destination Address — e.g. IO91SX. (obsolete).
                      • In AX.25 beacon text, with the [ APRS Data Type Identifier — e.g.
                      [IO91SX] (obsolete).

                      Grid squares may be in 6-character form (as above) or in the shortened
                      4-character form (e.g. IO91).

                      I actually raised this question a while back in the aprssig and was
                      told, again IIRC, that no popular APRS clients or sites actually
                      interpret and map stations using grid squares, which is a good reason
                      for WinLink to not use these formats. We were looking for a way to
                      shorten position packets from APRS Messenger on 30m HF.

                      Lynn (D) - KJ4ERJ - Author of APRSISCE for Windows Mobile and Win32
                    Your message has been successfully submitted and would be delivered to recipients shortly.