Onyx -- walking under own power for first time

[jwatte]

Babys first steps and all that. Walking under its own power. Raspberry Pi and ATmega32u4 controller on board. Linux/intel machine with Xbox 360 gamepad for steering. Gaits are based on some IK code I whipped together a few week-ends ago. Design, construction, and feet are all custom.

This thing is way too big! I'm going to have to do something about that huge center plate.

<a href="https://www.dropbox.com/s/qs2hnwnm7yyvdv4/2013-01-13 14.12.00.mp4">https://www.dropbox.com/s/qs2hnwnm7y...3 14.12.00.mp4

[Gertlex]

Great job! Very responsive and smooth. But yeah, that's a big-un.

[Th232]

Looks good, I presume the servos are having no problem carrying that load.

If you need to make it smaller, could you remount the tibia servos so that they're attached to the tibia by the side rather than the end of the servo? Aside from that, time to make a two storey chassis I suppose.

[jwatte]

I actually don't know a tibia from a femur -- is there an explanation somewhere?

The good thing with long legs is I get high speed. I'm thinking of mounting rubber balls on the ends of the legs to make it less noisy and save the floors a bit...
For the IK to be "perfect" I'd need to have a thin needle point toe, though. Perhaps I can use that in "combat mode" and use rubber while developing?

[Th232]

The terms come from general anatomy. In a leg, generally the femur is the long bone attached to the hip (or coxa), then the tibia is the bone after it. After that comes the tarsus, which isn't present so much in larger animals, but is found in a fair number of insects.

Servo-wise, the tibia servo is the one between the tibia and femur.

[Gertlex]

I've not thought too hard about it, but would a pair of wheels on each foot allow for the rotation that point contact would give? It could totally mess up movement, perhaps, though. I've not tested it, merely mused

[jwatte]

You'd need a servo to turn the wheels perpendicular to the direction you want the feet to "stick." If I'm doing that, I might as well do a 4DOF walker, where positioning the foot contact is pretty easy. I think I'd actually use a second sideways servo, rather than three downwards turning servos, for 4DOF. Maybe for next year :-)

[jwatte]

I grew tired of the 10 inch base plate size, and made something in the 5x5 inch size. The drawback is that I have to use several levels.
Bottom plate: Battery, hip servos.
Second plate: Power management, microcontroller, turret rotation servo.
Third plate: Protects electronics.
Fourth plate: A bracket, really. Guns and camera mount on this. Mounted on the pitch servo, which is mounted on the yaw servo.

Also, I'll strap the AD-11 PC on the belly of the beast (underneath the bottom plate.)

Here's a picture of different pieces in different stages of completion:

- Pocketing and profiling in aluminum stock.
- Filing and sandblasting.
- Matte black powder coat.

The big black tubes are my BB magazines. Today, I'm making the complex part that mates them to the gun chamber -- that'll be... probably not an immediate success :-)



Separately, I have the AD-11 and the 32u4 and a webcam and wireless networking all talking to each other, using a somewhat robust protocol. The IK engine runs at 120 Hz on the AD-11, and the webcam can give 30 fps when light conditions are good. CPU load on the dual-core AMD is about 80%, so I'm not sure I can fit autonomy into that... That's my backup for dropping connection -- it'll switch to autonomous. Robots with guns, thinking for themselves. Nothing could possibly go wrong with that ;-)

[jwatte]

Here's another update. I've been working on the hopper for feeding BBs to the guns, as well as the gun mount bracket which you can see not yet cut out or finished to the right in the previous picture.

Here's a magazine, plus the hopper that mounts to the magazine and the gun bracket, plus an agitator and a motor to drive the agitator. When holding it all together by hand and applying voltage to the motor, BBs come out at a fairly suitable rate. Here's hoping I won't have to aim upwards too much, as the slant on that magazone is only 15 degrees... I guess I could turn that agitator into an auger screw? ;-)



You can see the magazine on the left -- aluminum tubing, aluminum cap, powder coat.
The BBs are biodegradable 0.20g. I haven't tried these with my gearbox yet -- I may have to downgrade to the 0.12g; we'll see. The barrel assembly isn't done, so I can't test yet.
The hopper block is probably the most complex part I've ever made :-) It mounts a small electric motor, the actual magazine on the other end, and has a 75 degree angle cavity to lead BBs from the magazine into the nib you can see to the right, which mounts right above the chamber of the gun. Go go gadget: Ballnose Mill! Also, I had to make some custom soft jaws to fit the piece for the final operation, as it's milled on three sides and the final side is 15 degrees off square. That was exciting, too :-)
The agitator mounts to the motor with a 1.6mm set screw (the motor has a D-shaped 3mm shaft) and the motor itself screws into a cavity milled on the back of the block.
The block is only sandblasted so far, haven't coated it yet because I'm not sure of the fit and functionality -- I may have to modify the design and make another one, but so far, it's working surprisingly well!

[sarendt]

jwatte,

Really nice work, great detail in your posts, very informative

Looking forward to more! Do you have a design your aimming for as a final plan, or is it more adhock?

Cheers,
Scott