|
Post by donburch on Oct 1, 2006 11:13:21 GMT 10
Hi all,
On a couple of occasions Bones has suggested that we have a go at a group project ...
During Bones' stopover in Sydney I asked if he had any suggestion and he mentioned building a scorpion.
A Scorpion would be based on a six-legged walker, with the addition of two pincers (grippers) at the front, and a curved tail at the back. As such it could be built in stages, giving several milestones during the project.
Bones has obviously thought through construction of the tail, and while I got the general gist of it I'll leave it to him to explain properly; except to comment that the servos for up/down and sideways movement would be located in the body, and that he envisages the overall size being approx 600mm meter long by 300mm wide.
Personally I come from the software side, so it sounds pretty challenging for me - which is the best way for me to learn, so I'm definitely interested !
|
|
|
Post by Dingo on Oct 2, 2006 21:34:03 GMT 10
I think that is a great idea. Do we have a rough idea of the parts list so far? ie how many servos, what sort of microcontroller, what sensors make sense? We could also do bulk purchases to reduce cost (I'm thinking servos here )
|
|
|
Post by Robotman on Oct 3, 2006 15:51:33 GMT 10
Yesssss (that's a drawn out "yes") Certainly sounds like a sumo 'bot, and the tail will have a real sting in it for sure. (But what to do about an out of control robot slashing down upon the densely packed crowd?) I can see the tail being like a whip and I don't know how servos will cut the mustard for that. Maybe a high-torque motor pulling over a [very] short displacement? A long distance robot construction project is going to be a little difficult. Now, I'm not trying to be negative here - just the devil's advocate. Programming will need a simulator while the hardware is being built, or the hardware will have to built around the software - and neither will be free of bugs and does this mean making two sets of hardware and shipping one across the country because no software emulation can match the reality of loose/stiff joints, mechanical instability/flexing, high current drains, heaps of noise on the digital busses, etc? So maybe the hardware will have to be all built first, then shipped to a final assembly site, then the software written (probably at the site of final assembly) with maybe contributions by numerous code writers, but testing and bug reports made and written up by a single person (unless some of us relocate our homes) and, and, and... It will be a logistical exercise worthy of any decentralized commercial R+D department. And then my own construction interest is also probably consider a major "pie in the sky" by those who heard the very words from me's own mouth one recent early Saturday morning... Which leads me onto Well guys (Rod and Don), it was an interesting morning, if a little surreal at such an early hour. It's a real pity I couldn't stay longer - I would have liked to stay and discussed things a heck of a lot more. And cheers for the Servo Magazine thingy, Bones - I might have forgot to say so on the 'morn. But hey, don't let me stop anyone from doing anything! Cheers
|
|
|
Post by donburch on Oct 4, 2006 17:42:36 GMT 10
When Bones first mentioned it, I thought "my ghod, that's a huge project", but after a little discussion I realised that the body is a fairly standard six-legged walker, two grippers on the front should be fairly straightforward, leaving the tail...
Bones already has ideas for the tail, being multi-jointed (like a spine) but operated by a few servos in the base connected to wires which go up through the tail. I thought the tail would be a great vantage point for a camera, but Bones is keen to have a laser there to add a "sting".
Botman, I would suggest that the tail would be impressive even without whip fast operation, if that simplifies the project.
Long distance development will certainly be an issue to contend with. I would suggest that we build several of these, so that several people (or locations) can work on it simultaneously, and to avoid the possibility of one shipping accident destroying all that work. I would certainly be interested in building one here in Sydney, though we will need to do some planning and estimate cost before I can commit.
Thanks for the comments - even the devils advocate - and keep them coming.
|
|
|
Post by Bones on Oct 11, 2006 15:46:15 GMT 10
Hi Guys, Back from Bathurst. It was great. It was excellent to catch us with Rod and Don. We had chats on heaps of topics. I would mind making a scorpian. Here's the tail design. www.wolftronix.com/tail/index.htmI thing we could design an I can get route it out and send across in a flat pack. Probably plastic to start off with or maybe aluminium. We can bulk buy servos cheap and other parts. I think ultrasonics. laser in the tail that can range find as well. Cameras, pincers etc. I think I need to get a bug biik and get a closer look. Bones
|
|
|
Post by Bones on Oct 13, 2006 14:54:37 GMT 10
|
|
|
Post by Bones on Oct 14, 2006 15:44:45 GMT 10
|
|
|
Post by Dingo on Oct 17, 2006 19:39:12 GMT 10
I just get a 403 error on that link is all - then I go to the homepage and it is something to do with expression of the gay psyche.
Got a cached page etc?
|
|
|
Post by Dingo on Oct 17, 2006 19:42:00 GMT 10
One thing I think would be cool is if we could have our community robot designed so it could "play" with either beacons (IR or otherwise) or other identical robots.
Just a thought, so that 2 or 3 of these bots can interact in some fashion, even if it is just a very simple interaction or something as complex as following one another.
Since this sort of action can be greatly influenced by design, we should add it into the design melting pot.
|
|
|
Post by Bones on Oct 17, 2006 20:46:00 GMT 10
|
|
|
Post by Dingo on Oct 18, 2006 21:44:39 GMT 10
I like that a lot. I think the body will need to be wider to fit servos and batteries and microcontroller(s). The legs are probably the hardest part - I think that design has one too many joints. What sort of materials are we thinking of for the legs and feet? Just wire with rubber feet, rubber feet but cut lexan/other plastic etc, something else? I'm just going to assume that everyone wants to replace the scissor claws with gripping claws I think it would be great if we could leave a sizeable chunk of the body free for future addtions, so if someone wants to add GPS, there might be room enough for it.
|
|
|
Post by Robotman on Oct 19, 2006 14:10:42 GMT 10
There's a new toy in the shops that's being advertised on TV - it's a remote controlled dinsoaur thingy and it has a whipping tail and neck. Maybe buying the toy and disassembling it to see how they got the tail to move will be extremely instructive?
Building a few units will certainly help, and a community of creatures interacting together could lead to some super "higher order" behaviour - you know, the 1+1=3 type of thing.
What is going to be the primary motivation for this robot? I mean, what will the robot want to achieve with its gift of "life"? It can't just go about whacking its sting into anything that gets in its way - escpecially its brethren, so maybe some "food" robots should be built that are essentially soft-bodied domes on wheels that scurry around, trying to hide from/outrun the scorpions and then play dead for a while if they get stung. These tasty little morsels are ideal starter robots for the less experienced amongst us. Kind of like that little floor-scurrying robot seen in Star WarsTM. Or, these food robots could also have legs.
This concept is like a basic model of prey and predator, where the food (prey) robots consume grass (recharge at power-pad stations), and the scorpions (predator) robots feed off the prey. Maybe the scorpions can be recharged by stinging a prey and sucking out some [electric] juice? "Population" ebbs and flows will occur out of this when the scorpions become too greedy and "consume" all of their prey and cannot be recharged themselves because their prey cannot get to power-pads (to top up on juice) without being constantly hunted down. The scorpions could learn to "harvest" their food, and the food robots (Fobots?) won't be so scared to come out and play.
Maybe some ZigBeeTM (et al) RF modules will allow the scorpions and their food to have "voices" and "ears" just like real animals do, and use IR collision detection/avoidance initially, to which we can add vision later. The beacons will still serve to limit range of travel. So, tune in at 2.4 GHz and listen to the wild life...
|
|
|
Post by Dingo on Oct 19, 2006 19:33:45 GMT 10
Well I think the idea should be to have the behaviour as unconstrained as possible.
A lot comes down to the means of locomotion - gangly complicated legs are going to dictate a certain behaviour, quick moving wheels another.
If we try and concerntrate on appropriate sensors and appropriate locomotion, the behaviour then will just be a question of software and can hopefully generate complex actions from simple code.
With sensors like IR and enough room for things like Zigbee expansions, the basic versions should be flexible enough for everyone's tastes.
I think this project would do well to be a solution to locomotion (legs) and basic array of sensor (IR, Ultrasonic, contact etc etc) and a cool tail and leave the more complex software to the individual. Of course then everyone can share their code and we all benefit.
So what are everyone's thoughts on the leg design?
Overall speed needs to be considered, as well as complexity and what sort of environment we want it to roam around in (flat vs uneven etc)
|
|
|
Post by Bones on Oct 20, 2006 18:11:32 GMT 10
Hi Guys, I'm thinking that the Zigbee's is a great Idea because you can use them as as position sensors (Beacons) and communications. I'm thinking the legs should be alloy. I can mill them. As for joints, I think we should loose the one near the base of the foot. That means there is still 4 servos per leg. 2 at the body joint for back/forwards and up/down and another 2 joints on the leg. If we make the size right in proportions then we will fit all the servos in ok. You can buy the normal servos in the strength of the jumbo servos (9-15KG) but hey aren't cheap. That's 9Kg it can lift from 1 inch out from the control horn centre. Thats huge. A standard size servo is about 3-6Kg. The tail I want to whip around. How accurate do we want it? ? Throw some ultrasonics in there and IR distance sensors in the feet to measure the terrain height for walking on uneven surfaces. I like the idea of interaction. Real robot wars not that remote crap. A high voltage arc stinger to disable the other robot. Speed wise. Quick would be good. www.cim.mcgill.ca/~jasmith/scorpion.htmlThis guy has done a fair bit of work in Matlab on the walking. and I know a guy in perth that owns 20 or so real scorpion for us to get videos of. I also know a professor at one of the Uni's here that built a robot spider when he was at a NSW uni. I could ask him for some walking info. Let us know what you recon. Bones
|
|
|
Post by Dingo on Oct 21, 2006 12:45:36 GMT 10
This video might be of interest www.youtube.com/watch?v=gsCbvl5Bak8It appears to use 2 normal servos as hip joints and a smaller servo as the knee. As this is intended as a community effort I wonder if overall price should be a fairly large consideration. ie what should have priority for the community, stronger and more expensive servos or cheaper and less functionablity? It is just a matter of working out at what level the project should be "pitched".
|
|