To LUGNET HomepageTo LUGNET News HomepageTo LUGNET Guide Homepage
 Help on Searching
 
Post new message to lugnet.org.us.laflrcOpen lugnet.org.us.laflrc in your NNTP NewsreaderTo LUGNET News Traffic PageSign In (Members)
 Organizations / United States / LafLRC / 166
165  |  167
Subject: 
Re: latest results
Newsgroups: 
lugnet.org.us.laflrc
Date: 
Wed, 16 Nov 2005 16:08:25 GMT
Viewed: 
1247 times
  
   OK, I remember now. It's sort of pattern-matching with the pattern being
encoded by the position of the sensors instead of software (which, incidently,

1) The sensor geometry is usually determined by optomizing the line following,
so usually repositioning the sensor array means you aren't optimized for
following the line. It's a trade-off, and it may work.

2) You better come into a well-known zig-zag at *very* close to a head-on entry
angle - no wiggling allowed! Simple if there's a long (enough) straight prior to
the zig-zag, but it might imply keeping speeds low, which might *not* be a good
trade-off in this event.

These are things I just didn't have time to test.  My newest line following robot
"Twisted" was made to follow a line with six inch radius turns.  I spent a great
deal of time working on it, and have at least four or five scrap prototypes on my
working table to show for it.  I was trying to build a robot that could go over 3
ft/sec, which I did.

And, you know, at the speed it's going, a sensor will totally cross the
line in about 20 ms.

I also just realized if the line is under two of the sensors for any length
of time, then it goes to zero sensors, it's most likely in a zig-zag...  hmm.

Steve



Message has 1 Reply:
  Re: latest results
 
(...) I never took the time to think about this in those terms...hmmm...I will have to think about the impact of this to my approach... (...) Or a break in the line...either way, go straight... And, as a side note, I rebuilt Sniffer on Sunday and (...) (19 years ago, 16-Nov-05, to lugnet.org.us.laflrc, FTX)

Message is in Reply To:
  Re: latest results
 
(...) encoded by the position of the sensors instead of software (which, incidently, would make it very tunable at run-time - hold the sensors onto 12L axles projecting forward, for instance, so you can slide them back and forth). Hmm. That should (...) (19 years ago, 16-Nov-05, to lugnet.org.us.laflrc)

7 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
    

Custom Search

©2005 LUGNET. All rights reserved. - hosted by steinbruch.info GbR