headers
Steve Hassenplug | 2 May 2005 15:48

Re: Analysis of robot behaviours?

On Sun, May 1, 2005 12:42 pm, Johan Hjertman said:
> But first little info about the robots. There are two robots, one is the prey (3
> bump sensors, 1 rotation sensor) that just drive around and try to avoid things
> and it has a lamp on it. The other one is the hunter(2 bump sensors and 2 light
> sensors). It avoid things and try to follow the prey.
...
>
> My question is really how do I evaluate the robots behaviour after I´ve run the
> program, so I can see if it helps for example if I change the learning rate for
> the neural network or if I change the sensors position.

I would think you'll want to run some controlled tests.  For example, dis-connect
the motors on the prey, and see how long it takes the hunter to "catch" it.  Try
several positions, then make your changes, and try the positions again, to see how
that helps or hurts the result.  (takes more or less time)

Once you can see how changes affect the result, you'll be able to better understand
how to evaluate it's behavior.

Steve
Permalink | Reply | 1 comment |
headers
danny staple | 1 May 2005 23:29
Picon
Gravatar

Re: Pneumatic hexapod omni-Directional 2 (repost)


That is awesome. Could you put a vid of it in action. I love the style
- it is functional, but equally stylish - the way you have arranged
the cabling makes it look really something. I can really imagine that
as a spider tank model looking rather fantastic.

Could you do a pneumatic logic diagram of it? There are so many
connections there - I am pretty bewildered after trying to follow what
went where - I imagine it is segmented, and with a couple of sections
- each leg is basically the same in relation to the previous leg (or
another leg depending on the gait set up).

Anyway - thoroughly amazed at it!
Danny
--

-- 
http://orionrobots.co.uk - Build Robots

On 5/1/05, Kevin L. Clague <kevin_clague <at> yahoo.com> wrote:
> I uploaded some pictures of PhD 2, that some of you saw at BrickFest 2004.
> 
> PhD2 is a neat and tidy version of its predecessor Hex-363.
> 
> PhD2 can adjust the sweep direction of each of its legs, allowing to walk any
> of
> six directions (forward, forward right, forward left, backward, backward
> right,
> backward left), and can turn right and left.  It can even walk forward/reverse
> with one triple leg group, and turn with the other triple leg group.  PhD2 can
> also walk in place.
>
(Continue reading)

Permalink | Reply | 2 comments |
headers
Kevin L. Clague | 2 May 2005 18:02
Picon
Favicon

Re: Pneumatic hexapod omni-Directional 2 (repost)

In lugnet.technic, danny staple <orionrobots <at> gmail.com> wrote:
>
> That is awesome. Could you put a vid of it in action. I love the style
> - it is functional, but equally stylish - the way you have arranged
> the cabling makes it look really something. I can really imagine that
> as a spider tank model looking rather fantastic.
>
> Could you do a pneumatic logic diagram of it? There are so many
> connections there - I am pretty bewildered after trying to follow what
> went where - I imagine it is segmented, and with a couple of sections
> - each leg is basically the same in relation to the previous leg (or
> another leg depending on the gait set up).
>
> Anyway - thoroughly amazed at it!
> Danny

Hi Danny,
  Thanks for the feedback!  Before I dig up a circuit diagram, I'll show you the
timing diagram for my original hexapod, hex363:

http://www.brickshelf.com/cgi-bin/gallery.cgi?i=698017

  Hex363 can only walk forward, but has a hexagonal body.  It gets its name
based on the sequence of the feet and when they touch the ground.  Here is the
sequence:

  A.  Start with all six feet on the ground.
  B.  Lift three backward swept feet using large pistons B,C,D
  C.  Sweep the legs (grounded feet backward, up feet forward) using I,J,K,L,M,N
  D.  Place all feet on the ground using large pistons B,C,D
(Continue reading)

Permalink | Reply | 1 comment |
headers
Kevin L. Clague | 2 May 2005 18:04
Picon
Favicon

Re: Pneumatic hexapod omni-Directional 2 (repost)

<snip>

>
>   Here is a discussion on how I do my pneumatic sequence designs.... Circuit 7
> describes how I designed my pneumatic quadraped.  It is easily extended to make
> a hexapod or octopod walker.

Oops.... http://www.kclague.net/Sequencer/index.htm

>
> Kevin
Permalink | Reply | 0 comments |
headers
Thomas Chesney | 2 May 2005 23:11
Picon
Picon
Favicon

Re: Analysis of robot behaviours?

In lugnet.robotics, Steve Hassenplug wrote:
> On Sun, May 1, 2005 12:42 pm, Johan Hjertman said:
>> But first little info about the robots. There are two robots, one is the prey (3
>> bump sensors, 1 rotation sensor) that just drive around and try to avoid things
>> and it has a lamp on it. The other one is the hunter(2 bump sensors and 2 light
>> sensors). It avoid things and try to follow the prey. ...
>>
>> My question is really how do I evaluate the robots behaviour after I´ve run the
>> program, so I can see if it helps for example if I change the learning rate for
>> the neural network or if I change the sensors position.
>
> I would think you'll want to run some controlled tests.  For example, dis-connect
> the motors on the prey, and see how long it takes the hunter to "catch" it.  Try
> several positions, then make your changes, and try the positions again, to see how
> that helps or hurts the result.  (takes more or less time)
>
> Once you can see how changes affect the result, you'll be able to better understand
> how to evaluate it's behavior.
>
> Steve

Actually this is a bit of a minefield - the performance of your predator ANN
depends on your prey ANN and vice versa. You will have to run controlled tests
but I think the evaluation criteria is up to you - length of time to
intercept/avoid the prey/predator, efficiency of movement (a big deal in the
wild), whether the predator is getting the right time between 'feeds', which is
defined by you. Maybe you will have to run a series of tests. To evaulate the
predator, select and stick with the same ANN for the prey then prioritise the
predator criteria, assign each a weight and use this to give each ANN a score.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Dan Novy | 3 May 2005 01:55

Was: Some brainstorming needed, now...

Remember the thread about building a 3D printer, or rapid prototyper?  
Well,  some one did it. With Mechano, not Lego, but still very cool.

http://blog.rebang.com/index.php?p=101

and the winner as a material was....  HOT GLUE!!!   We were on to 
something!  Glad someone baked this one.

--

-- 
Dan Novy
Technical Supervisor
Flash Film Works
Permalink | Reply | 1 comment |
headers
Steve Baker | 3 May 2005 07:54

Re: Was: Some brainstorming needed, now...

Dan Novy wrote:
> Remember the thread about building a 3D printer, or rapid prototyper?  
> Well,  some one did it. With Mechano, not Lego, but still very cool.
> 
> http://blog.rebang.com/index.php?p=101
> 
> and the winner as a material was....  HOT GLUE!!!   We were on to 
> something!  Glad someone baked this one.

Woah there.  Be careful - if you read the full description:

1) It's not a general purpose machine - the guy only implemented
    a simple turntable and a system that lowers the turntable at
    a constant rate as it rotates.  All he can make is hollow
    cylinders.

2) He's quite clear on the fact that hot glue is a poor choice of
    material - the end results are blobby and bendy with a lot of
    oozing between layers.

3) Necessary cleanup of the machine after each run sounds like a
    major pain.  He talks about needing to wash everything down with
    boiling water - so this isn't going to be a Lego machine anytime
    soon!

However, it probably wouldn't be that hard to add another degree of
freedom to allow it to build other shapes (well, it would be difficult
with Meccano because they have no computer and no sensors).  He has
solved several interesting and tricky problems for a hot glue solution:
(Continue reading)

Permalink | Reply | 0 comments |
headers
Nitin Patil | 3 May 2005 18:05
Picon
Favicon

Motor Multiplexer for RCX

Mindsensors.com is pleased to announce the 4 Motor multiplexer based on I2C bus
for LEGO RCX. This multiplexer lets you connect 4 or more motors using RCX-I2C
interface developed my mindsensors.com. 
These motors take separate battery (up to 35 Volts) depending on the motor
voltage, and capable of 600 mA in each motor. It also supports 255 speeds and
variable breaking power for each motor.
for more info visit
http://www.mindsensors.com/index.php?module=pagemaster&PAGE_user_op=view_page&PAGE_id=41

and
www.mindsensors.com

Nitin
Permalink | Reply | 0 comments |
headers
Philippe Hurbain | 3 May 2005 20:42
Picon
Favicon

Rama, a tiny Killough platform

<http://philohome.com/rama/rama.htm<http://philohome.com/rama/rama_top-v.jpg>>

Inspired by the work of Leo Dorst and Markus Matern, I built this small (and
simple) <http://philohome.com/rama/rama.htm Killough platform>.

Photos, videos, building instructions are  <http://philohome.com/rama/rama.htm
here>

<http://www.philohome.com Philo>

:         Member of
:        <http://www.freelug.org <http://www.philohome.com/freeluglogo.gif>>
: [<http://www.freelug.org French Enthousiast LEGO User Group>]
Permalink | Reply | 0 comments |
headers
Andrew G. Meyer | 6 May 2005 00:23
Picon

Swarm?

Would it be possible to create a Lego swarm? What language would be most
versatile for this task? I'm thinking this would make a good introductory demo
for a robotics class I'm teaching.
Andrew Meyer
Permalink | Reply | 4 comments |

Gmane