Subject:
|
Re: hexapod walker
|
Newsgroups:
|
lugnet.robotics.handyboard
|
Date:
|
Fri, 8 Jan 1999 01:50:49 GMT
|
Original-From:
|
Mike Jones <michaelj@wt.net!nospam!>
|
Reply-To:
|
michaelj@wt.&Spamless&net
|
Viewed:
|
924 times
|
| |
|
|
I have been researching this exact topic for almost a year now. I started off
looking at the Handyboard do to its popularity and strong robotic application
design but ultimately I've decided to use one or more Atmel processors do to
their speed and variety of models. I am seriously looking at using one of their
20 pin versions to control each pair of legs where each leg has 3 degrees of
freedom instead of your thoughts of only 2 DOF. If you were to adopt a similar
approach of designing intelligent leg controllers this would move much of the
tedious (read "CPU hog") process off the Handyboard and then it could be a better
central brain and leg "coordinator."
Since it looks like you are near the beginning of your research and design I'll
share some of my personal views. The leg lift motors must be very strong but
not as fast as the swing motor. This is because in the fastest gait, an alternateing
tripod gait, each middle leg must support half of the weight of the robot and
this is cantilevered thus amplifying the weight. If you never plan to do an
alternating tripod gait then you can reduce the strength of the lift motor because
each side of the hexapod will always have 2 legs on the ground at the same time
and so must only support one forth the weight of the bot, BUT it will travel
at 1/3 the speed of the tripod gait.
I am opting to try to put all my actuators (motor) at the hip joint to reduce
the moments of inertia. This should let me move the legs somewhat quicker than
if some of the motors were located at the knee. This choice also increases the
weight that the lift motors must deal with so it is a trade-off.
Hobby style servos can get very expensive to match the speed and torque reqirements
of even a very small walker. In my case I think I can use 2 standard (42 oz.inch)
servos (cost $10 each for cheapos and $20 each for brandname ball bearing variety)
and 1 high torque (90-100 oz,inch costing about $60). If I want the high speed
versions (.12 sec for 60 degrees instead of .22 sec) it will almost double the
cost of the servos but will also double the walking speed. My original goal
was for an 18-22 inch hexapod to have a max speed of about 2 mph which equates
to 6 six inch steps every second! there are NO hobby style servos that can come
even close. So the next best option is about 1/3 my orginal goal and plan to
redesign the legs in the future using geared DC motors instead of hobby servos
which lets me gear them to fit the need and use really fast motors.
If you care to discuss this topic any further contact me directly since we may
be staying somewhat off topic, unless there is enough interest on this list
to continue and the powers that be, Fred(?), think it would be appropriate to
discuss this here. I am interested in communicating with anyone out there that
may be doing anything even remotely associated with this.
Mike
> Hi all
>
> Iam thinking about making a six legged walking robot using servos and
> wondered if anybody has already built one.
> I want to use two servos per leg one for lifting and another for the
> sweeping action so giving 12 in total. One thing i want to know is how do
> you control twelve servos? i have seen some serial servo controllers that
> control upto 256, is it possible to control them with just the handyboard.
> also does the HB have enough proccessing power to control all these servos
> and still have enough spare for obstical avoidance etc.
>
>
>
> Any thoughts, tips etc.
>
> Thanks Russ....
>
>
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
