(Short) Rambling #6 (thoughts on bots)

I was going thru the internet yesterday, and I found this awesome post on Realitypod that shows some really cool robots in action, most of them made by Boston Dynamics (man would that ever be a dream job! I think I would be happy sweeping the floor there!) It is truely worth a read (and a watch of all the videos embedded) I think my favourite bot they mention is RHex, that is one amazing hexapod, followed closely after by cheetah (almost half as fast as we are allowed to drive on the higway, but from a 4 legged robot, now THAT is impressive) anyway, a really good site to visit, great for inspiration (I think I might have to pull my lego out when I get home and try my hand at a lego RHex :) )

And after you are done watching the videos on that site, have a look at these ones:

Rambling #5 (thoughts on bots)

I was browsing Hackaday a little while ago, and I came across This homebrew guided rocket and it looked pretty cool. I have a few model rockets and appropriate engines for them, some micro servos, a micro magician microcontroler and a few one cell lipo batterys from some cheap RC helis. 9 and 10 DOF IMUs are cheap and so are single and 3 axis gyros on ebay, and after seing the results of their rocket, I think a gyro or 3 is deffinately necessary instead of an accelerometer, altho I might make use of one of those too (the micro magician has one on board) So I have, or can soon have all the parts necessary to try my hand at building one of those. My father has been encouraging me to try building things other than various takes on obsitical avoidance, and I think this could be cool. My rocket body is a bit smaller than the one they used, so I am contemplating straping on another two rocket engines to the base of my Mis(Guided)Sile to give me a bit more thrust, and either using rear-ejection parachutes, much like this:

or none at all, after all, this IS missile, and after having a look at their rocket after it's flight, I dont know how well it would survive even with a parachute...

I have also been thinking about building a propeller powered missile/airplane hybrid using a rear mounted brushless motor like This one swinging a large prop to give me huge amount of thrust, 4 oversized rear fins, sorta acting like really small wings, and probably four fins at the front for pich/yaw, much like the sidewinder

The difference with mine, is I will have 2 fins in the middle, linked together to give me roll control, thus enabling me to have only 3 servos on board (each controling two fins for each degree of freedom, simplifying control, and saving on weight) I think if I do some looking around I should be able to get a motor/esc/prop/battery combo that will give me better than 1.5 to 1 thrust to weight ratio for the whole thing, enabling me to have it take off vertically and fly for a bit longer than the estes rocket motors.

Also possibly coming down the pipe, project wise are:
1. fabbing up a better mounting bracket for Going Whole HOG, letting me take it to some higher speeds without worrying about it coming apart, and possibly automating the thing, which would make it my fastest obstical avoider to date (Lets see if those picaxes REALLY do run at 64Mhz...)

2. As my last post said, I got a new chassis for AutoSquared, so when I get some time, I will get it posted on LMR, right now all it does is avoid stuff, but with my compass and GPS in hand I hope to change that soon...

3. One of these:

Except in a blue PAPER! Origami shell. Not really sure if it will work, but it will be interesting weather it works or not. probably going to be just RC not a robot (someone find me a sensor capible of seeing thru blue paper, and I will make it autonomus....

4. The missiles I mentioned in this post hopefully....

5. Possibly that electric bike hub motor powered big outdoor robot (if that thing ever makes it into the real world I am totally mounting a laser sighted paintball gun or potato cannon on it. :)

6. A 2 DOF hexapod ( I have all the servos I need, and I do already have 2 and 4 legged bots under my belt, and now I need more legs.)

7. giving my quadcopter a new body, and a pan/tilt camera with live feed down to the ground. I will have a seperate tx/rx controling the pan/tilt so I can focus on flying and someone else can get some footage.

8. at some point I might finish up GPS (my other outdoor bot)


One last thing before I sign off, If you are an FBI agent or anything, and are adding me to your watch list because I mentioned making small missiles, do not worry about it, if they ever get made, I am not going to be doing anything devious with them, all they will ever amount to is a slightly stabilized/ angle holding projectiles, or funny looking RC airplanes. But feel free to keep dropping by, I appreciate the page views :)

Rambling #3 (thoughts on bots)

I recently purchased an Adafruit Ultimate GPS and a cheap ebay compass sensor in persuit of building a serious outdoor UGV. I started a while ago a largish outdoor platform baised on drill motors, you can have a look at that Here but I am incresingly having some doubts that it will ever be finished. I built the bot with two used drill motors, and two new ones, and after some testing, I found that one of the used ones was nearly burnt out, and sucked a HUGE amount of current, it still spins, generally is on it's way out. Normally I would have replaced it, and got on with life, but by the time I found out, I had made that nice steel shell, and riveted it together, so replacing is not an option, as I would have to tear apart the intire bot. Not going to happen. Lessons learned: always design to take apart. I knew this, but I guess I just had to learn it the hard way. Anyway, back to the rant, I think, no problem, I will just put one of my HUGE 35 Ah 12 v SLAs in the bot, but now I am worryed that my 1/6 RC truck wheels will not stand up to the weight. No problem, just buy some 6" lawnmower wheel, and attach those to the drill shafts. Not so fast, the lawnmower wheels have holes in the center that are too big. So here the bot sits, with no suitable wheels, or power supply. Doing a little thinking and I am debating moving to a bigger platform still, using This electric bike wheel (I already have one from building an electric bike, but that is a story for another time.) As a rear/front wheel on a trike (rear or front will depend on weather I am using a conventional or tadpole style trike.) Having some experence with this wheel, I know that the 26" will be a bit much for a robot that weighs a cupple hundred pounds, I do not want to hurt anything/body by crashing into them with a very heavy robot trike going 35 km/h, that would be bad. Anyway, I was thinking I could put either a 20" BMX wheel or a dirt bike wheel of some sort on the electric hub and use that as my main drive system.
The problem is steering, I have weighed a few options, including using two drill motors for the other wheels and having a free swinging "tail" with the bike wheel on it, varing the speed of the drill motors would cause the tail to turn one way or the other. Pros: steering would be simple (just like a differential drive robot), three wheel drive, simplistic mechanical design, very little to go wrong. Cons: lots of math to figure out the size of wheels I need to ballence the speeds of different motors, and figuring out how much to vary the speeds of all three motors.
Another option would be to have the bike wheel on a tail, but have the tail coupled to a motor, and a feedback sensor (probably a variable resistor) so I could actively steer the tail. Pros: fairly easy steering, no math to worry about. Cons: finding a cheap motor that is powerful enough (maybe a windsheild wiper motor from a old truck or something) mechanics for coupling the steering motor to the fork might get tricky, having to hack a servo into a REALLY powerful servo, a bit less sturdy, no three wheel drive.
I could also have kinda the same setup, but have the tail mounted to the chassis and have the non-driven wheels turn. Same pros and cons as above.

Food for thought. I will leave you with these pictures.

Rambling #2 (Thoughts on Bots)

What do you get when you cross this Gripper

 and a quadroped like this?

Well, I don't know, you could get a number of things, but, I propose that you could have the makings of a pretty sweet wall climbing robot. Just get 4 of those grippers, mount them on the feet of your quadroped, and alternate "ungriping" one foot while the other 3 are stuck, moving it forward a bit, planting it firmly on the surface, sucking all the air out agin and repeating for each leg. It would not work on smooth surfaces like glass, but deffenately on brick/concrete walls, and maybe even on drywall.

Rambling #1 (Thoughts on Bots)

I have been interested in mini-sumo for a while, and I was thinking about how I could make an (almost) unbeatible robot. I have come to the conclusion that competeveness comes down to a cupple important things: traction, low profile fronts (traction does no good if the other guy gets under you and ruins it all.) quick response time while searching, and just enough power to push the other guy out. Ideally, tank treads covering the whole 100sq.cm (or almost anyway) with a scoop that deploys down would be the best (have a look at "the beast") However, building something like that is out of the question for most hobbiests, so I have come up with what I think would be almost as good, namely a walker. Now before you call me crazy have a look at this pic, and see that the walker I propose is not your average biped.

As you can see, the walker has a shell that moves around a round inner piece in a oval fassion. To see a more detailed discription of how this works, have a look at this link. This bot was not built by me, but it explains the walking method rather well. Anyway, the two pieces (circle and shell) need to have the same area, so it has equal traction when it is up as down (the importantce of this is evedent if you have ever seen two of these things push each other around (I built a similer bot using the lego mindstorms NXT system a while ago to vs against my friend in sumo, his robot had rubber on the shell, but not in the middle, and he lost ground every time his shell lifted in the air.)) So a little math (and help from a math teacher) later, and I figured out the required diamiter of the circle, and drew a scale drawing:



 

The bot will have almost ( the walking method requires some blank space, as evedenced by the drawing ) half of the possible area on the ground at any time, and sometimes all of it, which, if covered with an approate high-traction material will give a ridculas amount of grip.

 

The centre circle will be driven by a gearmotor in order to turn the bot, it does not need to be overly powerful, but fast enough to spin the bot at a good enough rate to not let it get ambushed from behind. The walking mechanisim will be driven by a cam system much the same as "clunkers" but with a much smaller cam, the bot does not need to take big steps, nor go fast when pushing, and will be able to get by with a less torquey motor.

 

At the begining of the match, the front plate will fold down giving the bot a nice low profile wedge front, and will also be hinged as in the picture so that if a robot hits the side of my bot instead of dead on the walker will still possibly be able to get under by continuing to rotate. As for opponent detection, just about any method will work, be it mechanical, optical, or sonic.

 

I know, crazy, but I think it might work.....