Jon Biddell wrote:
>
> > 1) GPS receivers are not accurate enough.
>
> Untrue. Not military grade, granted, but my eMap is accurate to 5.0 metres with 7
>satellites, using either the inbuilt or external antennae. I have had instances where
>I have had 8 satellites and the accuracy increased to 3.0 metres, but these are the
>exception rather than the rule.
I think you have just supported my point exactly. It needs to be sub 1.0
metre in accuracy for bicycles. You need to think of situations where
multiple services are close together; e.g. road, path, road, drains etc.
The way around this is to time adjust the data collected. Either
subscribe to the service that uses JJJ, buy the adjust off LIC, or build
your own (just record the GPS reading at an exactly known spot and
adjust your readings with your recorded difference)
> > 2) GPS receivers are not reliable enough.
>
> Untrue. The eMap (and, to a certain extent, eTrek) from Garmin regularly take a
>pounding in Marvin the Urban Assault Vehicle, and get kicked around and dropped at
>home, and have NEVER failed to perform.
Well, one day in the urban assault vehicle is not really that reliable.
I'm thinking of the kilos of batteries that one is going to need to map
a weeks ride along the great dividing range. As I understand it, some of
the current models might just give 12 hours on a set of batteries.
> > 4) GPS receivers do not have the capacity (waypoints) = required laptop.
>
> Mostly true - although the eMap has lasted me from Bowen Mountain to Maitland and
>back - it all depends on your speed. Greater Speed = more distance between TRACK
>points (as opposed to WAYpoints)= greater distance.
Firstly, we are talking about mapping a road - not just recording enough
point to show which road you followed. It is cyclic point 2 & 4 problem.
You can boost your waypoint capacity by having a laptop pull in the
readings all the time, but then I've got to add a weeks supply of
batteries for the laptop.
>
> > 5) Cost - software - very expensive, or Grass
>
> There are several Linux and Evilware apps that are available gratis - see
>www.gpss.co.uk for a good Evilware one, and he's open to someone doing an X version
>of it.
Of the ones I've seen, none of them are anything more than Micky Mouse.
Either you are up for a full, expensive package, or it is something for
just one part. And you will see what is really needed below.
Okay, despite all the above, I believe it is do-able, now, for about
$2,500 with 2nd hand laptop, gps card and a few canabalised sensors -
you create the combined gps/data logger/bicycle computer. (GPS, time,
speed, cadence, inclinometer?, light?). For extra, you could also add a
digital camera, which could take occassional snaps of the left side
front wheel.
So, we have the bicyclist with bicycle kitted out. The system wakes up
and starts logging when they start wheeling their bicycle out the door.
They do their days riding/weeks trip/whatever.
Next time they hit the internet, their system sends in all the data.
First step is to correct the GPS readings - a simple awk script looks at
each line, decides which regional set of correction data to use, then
finds the closest matching time and adjusts the GPS figures.
Next is to convert the GPS figures into your storage figures; perhaps
GDA (new AMG).
Next is to relate the figures to the rest of your data. Anyone who has
seen data from various different agencies will tell you that very rarely
do they line up. This usually involves a person to use intelligence and
say that point on that bend is most probably this point on this corner
on our collected data - process the data - then check it to make sure it
fits.
Perhaps we can skip the above step if we can train all our
monkeys^h^h^h^hbicyclists to log certain points, e.g. we drive a steel
pin into the foot path outside their home and say enter #01# everytime
you cross this point, and we have #99# for a marked point at their work,
etc.
Eventually, you build up an enormous amount of data that you can process
to create a bicycle map.
What goes onto the map is a bit political - some people argue for KISS
and that means one line of varying colour.
At the other end is the six line monster (three different lines for
three different bits of information.
The inner line is the road itself (3 lanes, one lane, heavy traffic,
very low traffic, commuter peaks, hot mix, concrete slabs, gravel, mud,
etc)
The outer line is road edge condition and a vital safety information,
from dangerous because there is no shoulder and rock wall (raspberry
jam), or crash barrier (broken legs), or breakdown lane (nice, but
rough), nil (huge potholes) clearway, parked cars, etc. This is what
your digital camera is snapping.
The centre line could be grade & surface.
That is what you need to be able to do. Just collecting GPS reading is
not very useful. It is a bit like printing out all the linux
documentation - totally useless unless it is filed and index and stored
in a retreval system.
I'm not interested in a GPS fiddle thingy, I want a full system that
allows
>
> > 6) Cost - labour - who is going to pay the person to process the data?
I will add here that Adventure Cycling, the US version of Bicycle
Australia, is currently surveying US bicyclist to see if they want GPS
data, and the $64K questions is - how much would you pay for it? {:-)
--
Terry Collins {:-)}}} Ph(02) 4627 2186 Fax(02) 4628 7861
email: [EMAIL PROTECTED] www: http://www.woa.com.au
Wombat Outdoor Adventures <Bicycles, Books, Computers, GIS>
"People without trees are like fish without clean water"
--
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug
