DPRG mailing list archives: [DPRG] Long Haul Way Points

Mon Sep 11 09:58:15 CDT 2006

Hello DPRG list.

I have just found your organization on the net and joined (have not been to
a meeting yet).  I am coming into this mail thread in the middle of it so
this may be a little out of context, but my company designs and sells GPS
auto steering systems for agriculture and road construction.  I am the
designer and have had systems in the field for the last 5 years.  I have
tried about every GPS imaginable.  The golden rule of accuracy and
repeatability is "you get what you pay for".  

Our customers require, at a minimum, 4 inch short duration repeatability and
12 inch long duration repeatability.  This is impossible to get from any
sports receiver.  Base band GPS is only accurate to about 30 - 60 feet
(depending on single or dual channel).  Our experience with WAAS is that it
is better than the publicized specs, but has had some reliability problems
lately.  We have been seeing approx 24 inches of long duration repeatability
from WAAS.  The best repeatability, by far, is from the purchased correction
services.  Our products have completely gone with the purchased services
even though the receivers are over $5500 (at the low end).  The published
specs on these receivers are 12 inch long duration repeatability, but we
have been seeing closer to 6 inch.  All of these specs are statistical so
repeatability is a little bit relative to you level of pain.

If you want centimeter accuracy you can go with a dual channel RTK system
with a base station, but you are talking about big $$$$ (price is coming
down).  If any of you are interested, I would love to design an RTK system.
The price for purchased systems are totally inflated relative to the actual
hardware needed.

I hope that this may have helped.

New Member
John H. Harrison Jr
Harrison Ag Technologies
Johnh630 at comcast.net

-----Original Message-----
From: dprglist-bounces at dprg.org [mailto:dprglist-bounces at dprg.org] On Behalf
Of -Old Dog-
Sent: Sunday, September 10, 2006 1:31 PM
To: David P. Anderson; dprglist at dprg.org
Subject: Re: [DPRG] Long Haul Way Points

All:

The differences in GPS coordinates don't come as any
surprise to me. That's why I also described the course
in terms of headings, distance, and pictures. After
yesterday's meeting two potential competitors walked
the course with their GPS's, and noted discrepancies
between their readings. If they'd like to share their
results, I'd be interested in comparing them to what
I've already calculated. The maximum difference
between what yours and my readings are about 2/1000.
That doesn't sound like much; until you realize how
much distance a single degree covers.

What does surprise me is not being able to get a lock
at some key way points. When mapping the course, I
made sure I could get at least 5 satellite's per mark.
But I wanted this to be a contest of ingenuity, rather
than who has the best equipment. In that spirit, Eric
offered an interesting suggestion yesterday, when
another GPS had the results you mentioned: Activate
your robot far enough away from the building to get a
satellite lock prior to the start of the attempt. Then
direct the robot to the starting area, and use
odometry to compensate. There is nothing in the rules
about taking that approach.

> For those who might like to try another challenge
while we are at FiarPark,
> here are the 2 waypoints to go from the meadow next
to the Science Place
> to the meadow next to Texas Star Ferris wheel, which
must circumnavigate the
> Band Shell, as jBot demonstrated in a round trip a
few months ago:

I have no problem with this, provided it does not
disrupt the planned contest or distract from the
people running attempting to compete.

-Martin

--- "David P. Anderson" <dpa at io.isem.smu.edu> wrote:

> Howdy,
> 
> I went by Fair Park yesterday to verify Martin's
> waypoints and got the
> following.  John, you might try these out with
> Google Earth:
> 
> start:	N  32.77715
> 	W 096.76190
> 
> waypt1	N  32.77753
> 	W 096.76208
> 
> waypt2	N  32.77770
> 	W 096.76266
> 
> waypt3	N  32.77790
> 	W 096.76249
> 
> Note that these are specified in decimal degrees
> (hddd.ddddd) as are Martin's
> rather than degrees-minutes-seconds.
> 
> The first waypoint is in the shadow of the Science
> Place building on two sides
> and I was only able to get three satellites at that
> location, so it is not
> a very accurate measurement.  Likewise with waypt2,
> which is in the building
> shadow and also under a large tree.  
> 
> For those who might like to try another challenge
> while we are at FiarPark,
> here are the 2 waypoints to go from the meadow next
> to the Science Place
> to the meadow next to Texas Star Ferris wheel, which
> must circumnavigate the
> Band Shell, as jBot demonstrated in a round trip a
> few months ago:
> 
> start:	N   32.77634		// Water cover next to TSP
> 	W  096.76179
> 
> end:	N   32.77631		// In front of Ferris Wheel.
> 	W  096.75965
> 
> 
> A little more challenging is the route from the
> Science Place meadow, 
> across the entrances and parkinglots to the field
> behind the Natural
> History Museum:
> 
> start:	N  32.77636		// Water cover next to TSP
> 	W 096.76179
> 
> end:	N  32.77821		// Behind Nat'l Hist Museum
> 	W 096.76369
> 
> 
> For those without obstacle avoidance working, you
> might want to at least
> run a demo course.  Here is one laid out earlier
> this year. This was for
> those who might have a navigating robot but don't
> have obstacle avoidance
> working yet.  
> 
> It begins in front of the Cotton Bowl and has a
> second
> waypoint at a bend in the road, and a third point in
> front of the Aquarium where the path is blocked.
> 
> Plaza in front of Cotton Bowl (intersection of brick
> paths):
> 
> start:	N  32.77790 		// Cotton Bowl Plaza
> 	W 096.76112 
> 
> 
> waypt1	N  32.77707 		// Corner of Band Shell:
> 	W 096.76085 
> 
> 
> finish:	N  32.77654		// In front of Aquarium
> 	W 096.76042 
> 
> Here is an ariel view with the route marked out in
> red, from
> upper left to lower right:
> 
>
<http://geology.heroy.smu.edu/~dpa-www/robo/jbot/jbot2/fairpark2_x.jpg>
> 
> 
> best,
> dpa
> 
> 
> John wrote:
> > I plotted the way points with Google Earth.  I
> don't know what is ground
> > truth, that is Google Earth also has errors.  I
> was comparing them to
> > the map posted on the web.  
> > 
> > Start:  Was in the grass NNE of the round doom.
> > WP1:  Looked like it is on the edge of the
> sidewalk.
> > WP2:  Halfway between building and sidewalk, 8
> meters from the corner of
> > the building (to the southeast)
> > Finish:  Across 2d street in the parking lot 195
> meters from WP2.  This
> > point was also located here by TopoUSA.  Since
> TopoUSA doesn't show the
> > building (all of Fair Park is just a green blob) I
> couldn't use it to
> > verify the other points. 
> _______________________________________________
> DPRGlist mailing list
> DPRGlist at dprg.org
> http://list.dprg.org/mailman/listinfo/dprglist
> 

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
_______________________________________________
DPRGlist mailing list
DPRGlist at dprg.org
http://list.dprg.org/mailman/listinfo/dprglist