I would like to build a robot with the following initial characteristics:
- walking robot
- laptop based
- having different sensors
Living in Boston I have been looking for robot clubs,
but was not successfull so far. It would be great if anybody could point me
to any existing.
Also if one can help me with pointing to projects that I could study as a
base for building mechanical and control system of robot... A lot of
questions including how much it could cost, how and where find motors, etc.
Thanks in advance.
May I suggeest something lighter than a laptop? Also, if this is
your first robot project, I suggest building something smaller than
rolls just to learn the basics.
The RoboBRiX Adventure set is nice for that. Lego Mindstorms
is also very nice.
If you have your mind set on doing a larger robot at the
start, then here are my suggestions:
I'd start with a Systronix JStik (http://www.jstik.com /) if
I wanted a fairly powerful system, or a VIA Mini-ITX system
if I really wanted a PC-class machine (running Linux, of
Walking robots tend to be extremely expensive. Unless you're
very good you'll probably want a hexapod. You need at least two
degrees of freedom per leg (two motors per leg). This gives you
12 motors. Supporting a laptop or a VIA Mini-ITX board takes
a bit of power, so this will be fairly heavy.
I would estimate that you might be able to get motors for $50
to $300, depending on how good you are at this sort of thing.
For example, look at Lynxmotion: their rolling robots are *much*
cheaper than their walkers. Their high-end walker goes for over
$1000 with upgraded motors and it is too small for what you
I'm thinking (in the future) of building something based on
the Shelob legs (http://www.iso-bots.com /). Each leg has four
motors. I want to scale this up so that it would be
a hexapod or octopod with a 12"-18" center. I estimate that
motors would average $200, which would mean (for an octopod)
spending $6,000+ just on motors alone.
Whereas my next big project (Harvey) is based on wheelchair
motors and will hopefully be able to climb stairs will have
a total hardware cost of less than $1,000.
D. Jay Newman
this is a great starting point for me. I was thinking to use laptop for
robot's brain to avoid using wires between
robot and main computer. My intention was to use wireless network
between them. Let me look at your suggestions/links.
Are there any projects that use non-PC computers for robots with
It was in my pasture when I programmed PIC controllers and made
electronics boards, but now I am kind of lazy, please do not understand me
and would like to concentrate on programming robot behavior rather than
doing mostly mechanical and electrical work.
I am thinking to find a prototype of a robot or parts that can be
reproduced quicker and then do more programming.
For that I may need to see what other people done. I hope that somebody
will respond and send me a link to local (Boston or close) robotics club or
group if such exist.
Another reason for laptop was that multimedia devices could be just
plugged in without writing any code. I need to learn how non-PC based
systems would read and process multimedia.
Yes, hexapod looks really right thing for me now. So my plan is start
- hexapod mechanics
- estimate how big it should be in terms of size, weight and cost
- look for robot's brain: laptop or something else.
Hexapods are fun.
Processing multimedia on a non PC platform will present a challenge.
Putting a physical laptop on a Lynxmotion hexapod will probably not be
possible. Consider bluetooth.
If you look at the EH3-R, you will find a neat platform. I enjoyed building
it greatly, and programming it was fun.
I "let my code free" into the wild, and presently it is being improved upon
Mine is at http://www.bio-bot.com/lynxmotion
For PC based robotics, people are looking into things like the mini-itx and
PC-104 style platforms. This still leaves you needing some sort of I/O
capability, which people generally fill with a microcontroller.
Whats going to interface with the sensors on the robot ?
How are you going to hook them to a pc ?
for some kits including a walking bot have a look at
www.lynxmotion very nice easy to build kits.
The hardest thing is peeling the cover of the lexan.
For an easy to use board talk a look at www.bdmicros.com
I've used a few of Brians boards for various projects and work.
Uses the atmel atmega128
For windows can download avrstudio from atmel and winavr
(gcc port) from avrfreaks.com
For higher powered boards www.newmicros.com
Depending on what you need to do
have a choice between
dsp based (pod products) or philips arm7 based(arm products)
Choice of forth , isopod , c or asm for both pod and arm products.
Quite a few others make small arm based boards.
arm7 is a 32 bit micro, quite a few companies make different varieties of
If you really want to use a motherboard look at the
via mini-itx motherboards 17x17cm
or wait a bit until the
via nano itx boards are avilable 12cmx12cm
Even then your going to need something like one of the
servopod or isopods or something similar for io.
If you plan on using lots of sensors you may want to offload the
reading and direct control to a small micro mounted next to it
then communicate via can or i2c / spi depending on the number of sensors
to the main controller.Less processing work for the main controller.
Can get some cheap pre-assembled boards from www.futurlec.com
for microcontrollers like pic, avr, 8051 etc
Might help if I get the link right
Forgot to say you can use these with c , basic or asm
Brians got some example programs on his website.
avrfreaks www.avrfreaks.com has lots of projects and information on avr's
some projects that can be done with avr's
I've never used them for pcbs but have quite a few of their
pic and avr boards. Good for cheap quick oneoff projects.
Some of the cheapest dev boards you will find.
Can take up to four weeks depending if they have stock or not.
I've had good service from them.
Seen a few comments on their pcbs
Seem to be a mixed bag on the pcb service.
Hi Ek, as others have mentioned, building walkers tends to have a
steeper learning curve, plus involve a lot more expense, that wheeled
bots. I've been working on and off on mine for a couple of years so
far. Controlling 12-16 servos and/or motors on legs is a lot more work
than 2 motors on a wheeled bot. The other really big problem is that a
good-sized walker will have trouble lifting and carrying around its
own batteries, let alone a notebook computer. The bigger the walker,
the heavier the motors, gearing, and batteries need to be.
If you want to take a look at a lot of existing walkers, I have a pile
of links. The walking machine catalogs are a good place to start:
AFAIAC, Rbt Full at Berkeley is the best single source for info on
walking bugs+bots. Also, a lot of people find my projects page on
walking of interest.
- dan michaels
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.