Dream Big, Build Bigger

New Project Time!

2011-05-04T19:04:00.000-07:00

It's been quite some time since I've posted here, so I'll try to piece together the last year or so for me in a just a few words. Now, to sum up a year of college is quite an impossible task, but I'm going to try and do it anyway.

At the end of the summer, after what was probably hundreds of hours of thought and work, I handed my trusted robotic snake (Leonidas) off into the capable hands of a Robotics and Beyond founder, and said my goodbyes to InfoPlexus, my full time summer IT job. Then, on September 5th, about a week after every other one of my friends had left, I shipped straight up to the greatest city on the East coast (Boston, obviously), enrolled as a student at Northeastern University.

Let me start by saying that coming to Northeastern University was the single greatest decision I have ever made in my life. Initially, during the college selection process back in senior year of high school, I was pretty set on going to an all-tech school (ie, RIT, RPI, etc). I wanted to get as technical as I could. But then, one weekend, I went and visited my sister (who also goes to Northeastern), and had a sudden realization, that maybe I didn't want to put myself into such a hardcore engineering school. And what a good choice that was.

The best part about Northeastern (or at least what I've made of the engineering program there) is that you can make your experience there whatever you want. If you want to go and just be involved in your major-related classes, you can do that. But, there's always so much other stuff going on on campus that you can take advantage of. I know for me, the classes there weren't quite enough to keep me satisfied. This isn't to say that they weren't challenging (because some of them absolutely were) So, I began to look elsewhere for things to be involved with on campus. I became a member of the wireless club, a group devoted to amateur radio and circuit development/fabrication. I joined the FIRST Robotics team of college mentors for team 125: the NU-TRONS, and finally, became involved in research as a freshman working on developing programs in OpenCL to run in a multi-threaded environment (graphics cards mainly - hundreds of individual cores all working at once). Out of these three things, FIRST absolutely took up the most time. During build season, it meant being at the lab on weekdays from 6-9, and all day on weekends. Anyways, long story short, Northeastern has plenty of stuff going on (including classes) to keep an engineering mind busy.

Right now, I've been working on a new project, and it's come quite a ways in. This time, instead of ground based locomotion, I've been working in another field: flying. The last few weeks, I've been working a lot on a robotic tricopter. A helicopter, typically has one rotor with a set of blades on it. However, for a tricopter, as you can guess, it consists of three equally spaced rotors. By controlling the level of power flowing to each rotor, the yaw, pitch, and roll can be controlled by either artificial intelligence or directly from the user. Right now, the tricopter is already in the early flying stages (though not quite capable of steady flight yet), but its process is moving along fast.

Here's a picture of the current state of the tricopter, with more to come:

Basic Stamp Surgery

2010-05-18T18:05:00.000-07:00

I know it's been a while since I've updated, but between work, running our high school's musical, AP classes and AP tests, I have been working away at it, bit by bit.

These last few months have been the stages where I hit every problem that is possible to hit, like bad cabling, broken servos, unresponsive controller boards, broken drill presses, addressing the bytes sent over the serial connection for each individual servo, the whole programming challenge (of both the microcontroller and the computer program, which are written in two completely different languages), and god knows how much more.

Up until the other day, there had been been no smoke or fire.

However, that just changed. While I was running a diagnostic to see exactly how much amperage the servos were drawing, I had 2 of them stalled, and the whole rig was taking a total of about 4 amps. My mistake was that I had without realizing it, run this current through a conductor on the PCB of the Basic Stamp. It wasn't going through the voltage regulator or any other electronics, it was simply using the conductors on the board to send it out to the servos. Even with only 4 amps, this was too much, and the board went up in a nice cloud of smoke.

All was not lost. It was only a relatively small section of conductor (as soon as I saw smoke, I immediately cut all power to everything, if it had been on much longer, the heat would have gone into the actual IC controller chip, which would have been much worse.) So, with a little bit of probing with a multimeter, I found out exactly what had been obliterated, and with a little bit of creative soldering, I managed to put in a secondary conductor that completely bypasses out the broken one.

Here's some pictures of the fix:

Here's where the actual board got completely smoked.

RIP PCB

But is all hope lost?

I think not

All fixed and ready to go

More construction

2010-01-24T18:07:00.001-08:00

So the second shipment of servos arrived the other day, along with the generous donation of some sheet aluminium from the guys at Farrell Precision Metalworks in New Milford. They made everything so much easier on my end, and the quality of the the cuts was something I couldn't have come close to doing with a jigsaw.

Anyways, 2 of the servos were screwy, (one only had half range with slow updating, and one just had a mind of its own, and didn't like being told what to do.), so I will hopefully be able to ship them back tomorrow for new ones.

There are 2 more servos added on to the snake, and it really is starting to look like a snake, where you can see the pattern of metal running down. The longer it gets, the smoother the action will be, and the range of motion it will be able to have will get much wider.

Anyways, here's some updated pictures. Check out how it really is starting to look like a snake now, and not just a couple segments hooked together.

Here's a straight shot of it. Notice the pattern now?

Curved up on the floor.

Almost three and a half feet of snake, hanging in the air.

Bent up on the workbench. Sorry about the mess, although that's always the sign of a good project.

I didn't know snakes came in squares either.

Movement

2010-01-10T12:27:00.000-08:00

I've reached a new stage; where I'm letting the computer do all the thinking, instead of manipulating individual servos with a single joystick axis, I'm letting the computer do the thinking, and only using one axis on one joystick, and that's just for speed control/direction. However, for the first time, I have some videos up (sorry about the quality) to show a general idea of the basic movement gaits. These will smooth out over time.

A cool shot of the underside of the servo controller PCB.

All booted up and ready to go

A new addition, the Arduino micro controller. Too bad it's broken, and will need a new controller sometime.

A single joint

The whole thing.... so far.

And finally.... some video:

More and More, Bigger and Bigger

2009-12-27T12:56:00.000-08:00

I'm finally up to 4 servos, and 5 snake segments. Motion/programming-wise, I'm still just up to rolling on the ground, but I've only put a few hours into that, so there will definitely be more cool gaits to come. (think climbing stairs, climbing trees, etc) As of now, there really is no program that makes it slither or roll, rather, I have two joysticks (2 axises each) where each axis on the 2 joysticks controls a particular servo in the snake. This lets me get to fool around with it much more easily than writing code, debugging, etc. Once I get the hang of controlling the articulated motion with the joysticks (which is MUCH harder than it sounds) I'll be able to write a program that will do it much more efficiently and effectively.

I've fully moved myself off of the two 6V lantern batteries in parallel, and onto a salvaged computer power supply. Since the servo controller runs on 4.8-7 volts, I'm using the 5V rail (red wire) from the power supply, which provides up to 20 amps. Each servo draws about an amp for normal usage, more when stalled.

Again, I can't stress enough that none of this would have been possible without the help of Robotics and Beyond. Anyone or anyone who has kids who have any interest of robotics, engineering, electrical, or just building wicked cool stuff, should absolutely check out their summer program. As a past student and current mentor, I can say that it was a great way to introduce me to the field of engineering.

Again, here's some more updated pictures:

Here it is after pulling itself up vertically on just 2 servo motors. You'd never expect it, but these servos have about 21 kg (300ish oz) of torque. These guys are real finger crunchers.

In a rolling motion, notice the short (but existent) sin curve on both the X and Y axis.

Heres my end piece with power switches, plug, and even a switch guard for the main power toggle switch. See the cracks on the plexiglass in the bottom left? Moral of the story: don't run debugging programs on a concrete floor...

Here's the two joysticks that are currently being used to control everything.

And finally, my 'new' power supply. God bless the guy who thought of grounding the green rail to ground with a paper clip to turn on the PSU, you have made my life so much easier.

That's it for now, hopefully next post I'll be able to put up some videos of it in action.

More! More! More!

2009-12-19T12:32:00.000-08:00

So, not too much has changed design-wise, except for the addition of the electronics/control module to the snake in the last segment (the Plexiglas one). As of now, that only contains a servo controller, but in the future, it will have power switches, a microcontroller, and maybe even a bluetooth modem for wireless capabilities.

So, here's some more pictures:

Ready to be wired up

Electronics/control segment. No servo means more room for the controlling elements.

All bent up

So even though there's only 4 segments, it's possible to get it to move across a surface. As the number of segments goes up, it'll be able to do a lot more cool stuff. On a positive note though, my power supply is no longer a 6V lantern battery, and rather, a wall transformer. If the three amps that gives isn't enough, then the next step may be to salvage a computer power supply and use one of the 6V rails from that. It's not incredibly professional, but it works just as well as buying a regulated power supply just for this.

That's it for this time. Hopefully next time, I'll be able to put up some videos of it's motion.

Happy Holidays!

quick note to everyone

2009-12-18T10:46:00.000-08:00

For those of you who don't know what I'm designing and building, I'm working on building a robotic snake as an biomimicry internship for the Robotics and Beyond summer camp. If anyone's interested, check out their website:

http://www.roboticsandbeyond.com/

Anyways, back to what I was talking about before. What I'm working on building is a robotic
snake, capable of moving in both the X any Y directions. Some guys at Carnegie Mellon did the same thing (and also did an incredible job), and you can see their work at:

http://www.youtube.com/watch?v=T62E-_pQt3c
and
www.modsnake.com

Once finished, my snake should be very similar to that, with probably roughly the same number of degrees of freedom. Some more pictures should be up tonight or tomorrow.

Anyone with questions/comments, I'd love to hear from you.

Taking Shape!

2009-12-10T15:44:00.000-08:00

Finally, it's really starting to take shape and start to resemble a snake! A few updates from my last post include: an additional segment and an extra servo reinforcement that hooks the back of the servo into the rotating chassis, as well as the horn. This allows for the force/stress to pass through the snake through two points, and not just the one that had originally been planned. This addition greatly improves the structural integrity of the snake. So far, I have two joints in the X direction and one in the Y. After one more in the Y, I should be able to get some very basic motion patterns down, but nothing as good as the final product.

I'm getting to the point where I'm comfortable enough with the design that I'm getting ready to get a nice big sheet of aluminium for the chassis pieces, and weaning off the scraps laying around the garage. Also, I'm getting to the point where more servos are going to be

necessary, and not just the 4 I got to start with.

I'd like to make a quick thanks to Michael Morrissey and Robotics and Beyond for providing me with the chance to do something as cool as this. Without their help, I would never have been able to do this.

Now, here's some more photos of the status of the project so far:

A shot of the segments each twisted at 90 degrees so you can see the different aluminium chassis pieces connected together.

The segments angled at slight less angles than before.

Another 90 degree shot from a different angle.

All the segments in-line.

Another shot of everything straight together, and you can see how the different pieces interact.

Coming Together

2009-12-03T13:28:00.000-08:00

12/3/09

This project's no longer just thoughts and crude drawings scribbled out on napkins, or even the CAD sketches I've put together. As of now, something worthwhile exists. It's not too complex yet, seeing as I only have one joint of the snake flopping around on my desk, but, it's there, it's moving, and it's working just as planned. It shows that it IS all possible, and I expect to have up to 4 sections put together within the upcoming week. Here are a few pictures of the work so far:

Here's a preliminary shot of the first round of work, with just one servo and one chassis mount

Here's another shot of the second servo mounted into place, with the top chassis mounted in too.

A shot from the side

Here's a shot of the two servos without their top chassis piece holding them together.

blast off

2009-11-22T18:16:00.000-08:00

So, the snake idea has been starting to take shape, despite the fact that it still looks nothing like one, and won't for a while at least. However, I did manage to overcome my first expected problem: getting a live data feed from the computer, through the basic stamp microcontroller and to the servo controller, in turn moving the servos. It's not a 2 independent axis sine waves working together to coordinate a snakes movement like it will be in the future, but rather, a single servo with 2 positions, chosen by whether the the mouse button is depressed or not. However, the hard part has been taken out of the way, and the more complex multiple servo control should be relatively easy.

Hardware wise, I've just come out of the crude "folded index card mockup stage", and have entered the stage of using real materials, including aluminium, plexiglass, and lots and lots of screws. It still resembles no snake you have ever seen, because it's literally just part of one of the many segments the finished product will have.

I've included a few pictures to document my progress:

A Close up of all the electronics powered up and ready to work

Circuit boards, power sources, and servos

Just the electronic equipment, before being hooked up

A quick view of just some of the code used for the most basic motor functions.

A servo and its aluminium mount: just a small part of one of the many sections of the entire snake.