Subject:
|
Wriggly Walker -- a simple steerable hexapod
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 15 Jan 1999 22:25:33 GMT
|
Original-From:
|
Simon Brooke <simon@jasmine.#StopSpammers#org.uk>
|
Viewed:
|
907 times
|
| |
|
|
I'm nearly at the stage of having a walking torbot; which is to say that
the chassis and mechanics are finished and working, but the steering
still needs work.
The general strategy is very simple, and works well.
The machine has three transverse axles, each driven by a worm drive from
a single longitudinal axle, and each driving a leg pair. This mechanism
keeps all three leg pairs in sync. The front leg pair is on an
articulated module, and the longitudinal axle has a universal joint in
it at the point of articulation.
Each transverse axle has a crank at each end, and the top of a leg is
attached to the outer end of each crank. The leg then goes down through
a pivoting guide situated vertically below the transverse axle. Finally
each leg has quite a large foot. A cross sectional view through a leg
pair mechanism therefore looks like:
++
| ||
+--+--+|
| | |
++ O |
| |
#- -#
| ====
|
|
|
====
The centre pair of legs is 180 degrees out of sync with the other two,
making for a tripod leg arrangement on the ground at any time.
The legs are length 8 axles; the transverse axle is length 8; the crank
arms are the length three plates with cruciform end holes and a circular
mid hole (the length four purpose designed cranks might work better but
were rejected because it scales the whole model up beyond the parts I
have available). The drive gear is a 16 tooth gear.
The longitudinal axle is belt driven by a single motor, mounted in the
middle of the model between the central pair of legs. In earlier
prototypes the motor was at the rear end of the model, but it was found
that it is critical to a good walking action that the centre of gravity
be as near the centre of the model -- and as low -- as possible.
Without the RCX unit the chassis performs well, coping with substantial
irregularities of terrain and turning nicely. Unfortunately, with the
leg spacing I've got the only place to put the RCX unit is on top of the
model, and with it in place performance is substantially poorer, parlty
because the weight is at the mechanical limits of what the model can
stand (the legs flex visibly) and partly because the centre of gravity
is too high leading to instability on anything but a level surface.
There are also problems with cable length. I haven't yet found a way of
mounting antennae which are sufficiently light weight not to upset the
balance of the model close enough to the RCX for the short length
cables, and I'm already using one of the long cables for the motor. The
solution is probably to get hold of a couple more long cables...
The final problem is steering. Currently a dual-acting pneumatic ram
allows the front-leg module to be turned relative to the main module,
allowing the model to turn. However, I don't yet have a way either of
powering the ram or controlling it. I have a pump and a couple of
micro-motors on order which may sove this problem, and I am working on a
system for using the light sensor to detect when the steering is centred
(I want to keep the touch sensors for collision avoidance antennae).
The only part vital to making this model which isn't in the Mindstorms
kit is the universal joint; the pneumatic steering isn't either, of
course, but (a) it doesn't work yet and (b) there are probably simpler
solutions I haven't thought of.
Beam legs would probably be better than axle legs but
(i) I don't have enough long enough beams and
(ii) I haven't het worked out a simple way of making a swivelling leg
guide
which would work for beams.
Photos on my web site later,
--
simon@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/
-- mens vacua in medio vacuo --
--
Did you check the web site first?: http://www.crynwr.com/lego-robotics
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
