I am looking for a ready-to-go (and ready to buy) solution for getting
precise (+/- 2 inches) local positioning information into an
application running on a laptop computer. The system must cover about
a square half mile of hilly terain. The area is open fields, but there
is not continous line of sight. This is for a mobile robot
application.
I need about 10 samples per second. The max speed of the robot is
about 15mph.
I've looked into differential GPS systems, but they do not seem to
have the acuracy I'm looking for. I'm open minded though, if there is
a solution I missed.
Hopefully the solution will be <$5,000 total cost including any base
station. |
Request for Question Clarification by
mathtalk-ga
on
03 Sep 2003 20:37 PDT
Hi, bigjosh-ga:
I think you are right in suspecting that commercial GPS units will not
give the required accuracy of position. I believe that in warehouse
applications one finds successful designs which combine inertial
measurement units (IMUs) with "known" station information, often in
the form of bar-coded reflective strips. In this way cumulative
errors in the IMU data can be compensated for.
I'd love to hear more about your intended application and to explore
the possibility of using IMUs with some novel "fixed station" device
suitable to the open field/hilly terrain described in your question.
regards, mathtalk-ga
|
Request for Question Clarification by
mathtalk-ga
on
10 Sep 2003 08:27 PDT
Hi, bigjosh-ga:
I'm not sure if you saw my original request for clarification.
Arimathea-ga has posted a Comment below, expressing his interest in
your question as well.
If you could give a better idea of the intended application, it would
help to create confidence that we can deliver a truly useful response.
For example, you describe the terrain as hilly, "open fields, without
a continuous line of sight. How much preparation and measurement of
the site would you be able to do in installing the application? If
the $5,000 figure is going to cover the costs of this effort as well
as that of any hardware/software, it will be a real challenge to cover
an area of the size you suggest (a square half mile down to a scale of
2 inches).
This is not to say it cannot be down, but to suggest why further input
from you would be helpful in providing a workable answer.
regards, mathtalk-ga
|
Clarification of Question by
bigjosh-ga
on
10 Sep 2003 16:27 PDT
I am trying to build a gas-powered robotic lawn mower to replace my
current electric one.
The current one uses a semi-random mow pattern combined with burried
permiter wires. While this works well for small plots (<5 acres), it
is not suitable for the larger plots (20-100 acres) that I want to be
able to handle.
For ease of implementation, I'd prefer absolute co-ordinate data to
inertial/dead-reconned position data. I've done
gyro/compass/accelerameter stuff before and it is just too much work
and too fault-prone. There are enough other hard parts to this project
I'd rather focus on. :)
I am certainly willing to do some site prep to get this to work.
Idealy, I'd like to get away with just installing 1 or 3 base
stations, depending on the technology. I'd like to avoid doing any
more digging or large-scale wire laying. It just takes too long and is
very fault prone.
I've seen laser and IR beacon systems, but I don't think these will
work becuase there are hills that obstruct the view between parts of
some of the plots. They also seem unreliable.
I've seen DGPS-based and even 802.11-based systems that provide the
level of service I need, but all the ones I've found seem to be
experimental or hobbyist. I'd like a ready-to-buy system that I could
easily integrate without having to do much work.
I feel like this is well within the capabilites of current technology-
I'm hoping it exists already and i just can't find it.
thanks.
|
