Subject:
|
Re: How many people signed up for the NXT Developer's Program?
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Wed, 18 Jan 2006 22:33:43 GMT
|
Viewed:
|
2044 times
|
| |
|
|
In lugnet.robotics, John Brost wrote:
> I have a great deal of experience with GPS. The accuracy you mention is indeed
> correct (the GPS I'm looking at right now is specd at 5 meters horizontal
> accuracy). However, when you look at second-by-second accuracy, it is much
> better. A great deal of the error in many GPS receivers is caused by drift,
> which is how much the reported position of a stationary object moves over time,
> usually over the course of several hours. When mobile, the drift isn't so much
> of a problem, unless you are trying to drive in the same circle for hours on
> end. The GPS unit I'm specifically thinking about using is one of the better
> ones for this sort of thing, and given a clear view of the sky, I think it would
> do okay. I've been impressed with its performance so far in other applications.
> I could even do a running average on the reported coordinates to clean up any
> errors. Your concerns are indeed valid, but my experience and intuition tell me
> it can be done.
>
> Obviously I'm not going to have something that is highly accurate, but I am
> pretty confident that I could get something to follow a course within 3 feet or
> so.
>
> As for a good use for such a robot, well I never said I had a good use for it,
> besides the fact that I think it would be cool to do! :) After all, why do
> people climb Mt. Everest? Because it is there.
>
> Maybe navigating autonomously to pre-programmed points, sort of like outdoor
> line following without the line. If I got really ambitious, I did find this
> event called Robo-Magellen, which appears to be an autonomous outdoor robot
> competition similar to a mini-DARPA Grand Challenge.
>
> http://www.robothon.org/robothon/challenge.php
>
> Could be interesting, but it would need more than just a GPS for guidance, maybe
> the NXT's ultrasonic detector for obstacle avoidance.
I totally agree. Try playing with GPS 'cos now you'll be able to! The RCX was so
limited by only being able to read a single dimensional value with an effective
accuracy of about 1%.
I've played with GPS a fair amount too and your description of drift is pretty
much dead on. Caused by varying atmospheric conditions affecting propagation and
local multipath effects amongst others. The other more disappointing problem
I've seen is the blend when one satellite goes out of view and another one comes
into view. You often get a position (or track) discontinuity as that occurs.
We've experimented by augmenting GPS with other sensors which know perfectly
well that a sudden 10 foot move to the left, for example, never in fact really
happened, and thus we can maintain a GPS bias value.
The possibility that a small army of robotics enthusiasts will now be able to
apply themselves to experimenting with this kind of thing is fantastic!!
JB
|
|
Message has 1 Reply:
Message is in Reply To:
77 Messages in This Thread:
 
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
