DPRG List  

DPRG: Sonar users: share your ideas!

Subject: DPRG: Sonar users: share your ideas!
From: Daniel Herrington Daniel.Herrington at nsc.com
Date: Mon Mar 8 11:43:32 CST 1999


It may be that less of the signal is being reflected by the 12" wall
than by the kitchen (> 12") wall.  The sonar "beam" is a cone roughly
30 degrees wide, some of which probably passes over the 12" wall,
depending on how close the robot is to the wall, and how high up the
sonar transducer is from the floor.  Maybe an experiment would be to
double the height of the wall to 24", and test the sensitivity again.
I know that SR04's sonar units are pretty low to the ground, at least
in comparison to the height of the contest walls.  Maybe Trilobot's
are too high up for the same walls.



=> Replying to Ralph Tenny's message, "DPRG: Sonar users: share your ideas!" (Mar 8):
 > After the contest, many of us gathered at Spaghetti Warehouse to unwind. I
 > was lucky enough to sit with Jim Brown, and we discussed the puzzling
 > behavior of Trilobot, which had demonstrated problems identifying the wall
 > without hitting it.
 > It developed that an easily-fixed mechanical problem prevented scanning at
 > right angles to the for/aft axis. In Jim's kitchen, everything worked
 > well -- Trilobot was able to identify the wall (kitchen cabinets) and turn
 > to run parallel with it. At the contest, turns were not completed; it
 > looked as if the Trilobot could no longer see the wall and turned back,
 > hitting the wall.
 > So, here's the deal: Have any of you sonar users noticed apparent loss of
 > sensitivity or accuracy in scanning hard surfaces like kitchen cabinets as
 > opposed to softer surfaces like our styrofoam walls?
 > David Anderson's SRO4 uses sonar to locate the cans, but a lower relative
 > return from the walls should only make the cans stand out more clearly. 
 > Do any of you sonar users have any anecdotal or test edvidence to add to
 > our knowledge base?
 > Ralph

- -- 
Daniel Herrington
Product Development Engineer
National Semiconductor Corp.
email: Daniel.Herrington at nsc.com
phone: (817)468-6684


More information about the DPRG mailing list