Subject:
|
Re: LEGO MindStorms 2.0: Alternative transmission techniques
|
Newsgroups:
|
lugnet.robotics.rcx
|
Date:
|
Fri, 6 Dec 2002 17:24:29 GMT
|
Viewed:
|
3287 times
|
| |
|
|
Hi, I know that the original post (read below) is kind of old, but the
hole project seems quite interesting. I also read the reply that Mario
posted. I agree with him in all the points. I'll assume that you want to
develop a vehicle, not a robot with legs (just the implementation over the
vehicle is enough complicated, trying to implement it over a two legs based
robot would be practically impossible using just one life)
But I do think you can implement this project by stages. This supposes that
you don't have to implement all the milestones of your project at the same.
As Mario pointed, there're several complications, such as navigation
control, wireless implementation of the communication, etc.
Well, if you prioritize your project, you should came to something like
this.
1. The first thing you should solve is the navigation algorithm/control.
2. Intelligent self-caring sense of the robot
2. wireless extension of the robot
4. other improvements, such as intelligent decision making etc.
In order to start with point one, you must have some temporary solution
to point 3. This means you should use some wired implementation, such as an
I2C (inter integrated circuit) controller network, that can spread (and
recieve the responses of the RCX) the IR protocol over an extensive area
(in order to test the robot).
Then, using rotation sensors in order to count (in a more accurated way)
steps from the motors, due to the fact that their power varies according to
batteries; you could implement the navigation control. I've seen on the net
radar sensors (I don't remember where, but you can search over the
newsgroup) so that could help to avoid obstacles.
Once the robot can navigate acceptably, you can focus on the wireless
solution, maybe using some RF controller (such as an RF PICMicro MCU (I
don't know if they have been lauched yet, but if not, soon they'll be) or
another controller family)
Once the wireless solution is working fine, then you work on another
level of abstraction (such as a fuzzy logic decision making method) in order
to make the robot "think" over the enviroment that sorrunds it.
Well these are my thoughts.
CL.
"Mario Ferrari" <mario.ferrari@edis.it> wrote in message
news:Xns92BC916D95D8Fmarioferrariedisit@209.68.63.236...
> Hi Andreas
>
> Your project requires you face some very tough problems. I'd say that you
> cannot reach your goal, unless you add a lot of non-Lego hardware stuff
> to your robot and your building.
>
> IR communication is not appropriate, as it requires the robot and the
> tower are always in visual contact each other. Thus you should look for
> some custom radio interface.
>
> However, the biggest problem is navigation. There's no easy and reliable
> solution that allows your robot to know precisely where it is inside an
> area as wide as a building. It could use some measurement system (like
> monitoring the distance covered by the wheels) to have a rough idea of
> its position, but unfortunately this technique acumulates errors in a
> pretty fast manner and leads to wrong assumptions after some movements of
> just a few meters. Other tecniques require external reference points,
> which the robot can use to deduce its position: this are either
> artificial landmarks like lines or spots on the floor, or some kind of
> radio or light beams. All of these require complex preparation even to
> make the robot effectively navigate a single room, the work to prepare a
> whole building would be huge.
>
> You have choose quite a complex matter for your project :-) It's been
> discussed many times on the lugnet.robotics newsgroups. If you want to
> know more I suggest you search lugnet.robotics for "navigation", you'll
> find some interesting (and long) discussions.
>
> HTH
>
> Ciao
> Mario
>
> "Andreas Nauerz" <a0355137@addcom.de> wrote:
>
> > Hi,
> >
> > we want to build a Roboter with the following characteristics and
> > under the following conditions:
> >
> > The roboter should be able to drive through a whole building by
> > itself. During his drive he detects obstacles, walls etc.
> > Everything he detects he sents back to a computer.
> >
> > The computer is the brain of this application.
> > He holds the program (the logic) of the whole application.
> > He gives the comand to the robot where to drive and accepts the
> > feedback from the robot.
> > Depending on what the robot sends he sketches a plan of the building.
> >
> > So this roboter is controlled by a PC and acts as a little architect
> > ;-) We want to realize this via programming the robot with Java, e.g.
> > a Java RCX API.
> >
> > First of all... is this whole project possible ?
> >
> > Our problem (!):
> > What happens if the robot leave the room ?
> > How can we still send data from the PC to the Robot (for control of
> > the movement) ?
> > How can we still receive data from the robot to the PC (for feedback
> > about obstacles and walls) ?
> >
> > We do not want to use wires of course.
> >
> > Any ideas ?
> > I guess IR is inappropriate in this case, or ?
> > Are there other wireless solutions ?
> >
> > Thanks a lot for your support !
> >
> > Regards, Andreas
> >
> >
> >
> >
>
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
