To LUGNET HomepageTo LUGNET News HomepageTo LUGNET Guide Homepage
 Help on Searching
 
Post new message to lugnet.roboticsOpen lugnet.robotics in your NNTP NewsreaderTo LUGNET News Traffic PageSign In (Members)
 Robotics / 11757
11756  |  11758
Subject: 
Re: Extra RAM
Newsgroups: 
lugnet.robotics
Date: 
Thu, 22 Jun 2000 15:27:18 GMT
Viewed: 
675 times
  
John Barnes wrote:

I feel compelled to add that using a bit map greatly simplifies the
code involved in handling a map. And it allows mapped features to
be classified easily too. Each byte, in a 2D "map" array represents a
patch in the real world (a paxel? or loxel perhaps) and it can take on
different values depending on whether a location is uninvestigated,
known and off limits, known and within limits, contains a moveable
object etc etc. It is quite hard to represent this richness using
vectors. But as has been noted before, the RCX's 32k makes it a very
tight fit for these types of applications.

This really depends on the model you use. If you think of a simple
vector map, which only encodes spatial data, you are right. But think of
an model, which is divided into to major parts: spatial data and object
data. The spatial data describes the outline and position in space.

+---------------+           +----------------+
| Object 1      | <-------> |  Spatial info  |
+---------------+           +----------------+
        ^                           ^
        |                           |
        v                           v
+---------------+           +-----------------+
| Object 2      | <-------> | Spatial Info    |
+---------------+           +-----------------+

a.s.o
The full list of all spatial information will describe the 'world' in
means of position. While the list of Objects describes the 'world' in
means of Objects like tables, doors etc. If you want your robot to go to
door 4, it only have to scan the list of objects to find the Object
'Door 4' check the link to the spatial data and than it can calculate
the 'best' way to go there.

Navigating in such a approach is a little bit more complicated than in a
loxel map. But not much more. Think of the definition of an virtual
object 'Fairway', which outlines describes the area which is normally
movable by the robot. Even the Robot itself can be represented inside
the map by an the object 'Robot', which is movable. inside the spatial
definitions of the object 'Fairway'. Also movable objects (like chairs
a.s.o) which resides inside the 'Fairway' can be found with simple
algorithms.

Even with such an approach you have more flexibility than with a loxel
array. The downside of this approach is, that you need more
computational power.

Regards Carsten



Message is in Reply To:
  Re: Extra RAM
 
I feel compelled to add that using a bit map greatly simplifies the code involved in handling a map. And it allows mapped features to be classified easily too. Each byte, in a 2D "map" array represents a patch in the real world (a paxel? or loxel (...) (24 years ago, 22-Jun-00, to lugnet.robotics)

2 Messages in This Thread:

Entire Thread on One Page:
Nested:  All | Brief | Compact | Dots
Linear:  All | Brief | Compact
    

Custom Search

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