Monday, January 18, 2010

Something is obvious

The control system must use polar coordinates. All wheels will only go one direction that way. Hm. Seems right.

Sunday, January 17, 2010

Moving forwards

Progress was slow today. Had issues with glue and wheels. Thinking about picking up and raiding an RC car for parts. I'll need a controller, but twice the motors, so not sure if that'll work. Have a weird mental block on hooking up the controller to the motor and stuff. Slowly working through it.

Saturday, January 16, 2010

Real progress today

I realized that the easy hands-off stuff is waiting on hands on stuff to move forward. It's easy to say you'll make a little progress every day when there are little things to do. I have big things that need to be done, and a lot to be figured out, so the little incremental daily bits aren't working.

Today I made some real progress towards robots. I have the basic structure put together now. I'm realizing I need different wheels, as the ones I have are not going to work. I'll also need to figure out how to mount the motors. This isn't the biggest challenge for me right now.

What I need to figure out next is figuring out how to drive the thing.

I have:
Motors, TReX Jr. Motor controllers, a mega arduino. Nothing is connected to anything else. I'm a little overwhelmed at the moment as to where to begin. I guess I can start with the motor controllers. I suppose moving on to the Mega Arduino would be best once I can figure out what it needs to output. Then I guess I can make a little sequence for it to get the motors to move in a certain sync relationship.


After I get new wheels, I need to figure out how to pass signals down to the motor. Then once I have the arduino generating patterns, I can call it done for the weekend.

Next up after all this is done?

Coming up with ways to map intended motions to motor instructions.
Finding some way to pass this guy instructions for things.


First thing, I need to figure out how to hoook u

Tuesday, January 12, 2010

Not doing so well...

Bit of job-drama yesterday enough to occupy thoughts on non-robot things. By the end of the week, I think things will change. In the mean time, not sure how much I will be able to focus here.

Monday, January 11, 2010

Frustrated

Slept most of the day today, so made no progress. Being sickly right now does not help forward motion at all. Need to do more tomorrow.

Saturday, January 9, 2010

Occupied?

I studied for and passed the amateur radio technician license today. I haven't had my mind on much with this except for a few things:
- I need a drawing of this before I get too much more into it.
This poses a conundrum as I can't draw and I'm not familiar with any CAD program. I'll try to put something together in Blender tonight/tomorrow and see what comes of it. I'm told Blender has amazing properties. We'll see.

- What kind of communications am I going to need?

I'd toyed with the idea of creating a communications protocol for these guys. That might be next year's project, though.

Friday, January 8, 2010

Thoughts before breakfast

So I started out thinking this morning about the rotary electrical unions. In thinking about it, I feel it'd be a bad idea to buy these. They're extremely hard to find and very expensive when you do. Most of them are made for high power environments (such as generators) so are way bigger than we'd need. I spent a bit of time talking to Carl Pisaturo about his rotary unions at Dorkbot a while back. Although his are very pretty, I think we could come up with something a lot less polished that would do the job slightly more off the shelf, but following about the same concept of rings and combs. These could be constructed as "caps" that go on the tops of each of the legs. I do think more thinking is required in the comb part. I'm not convinced of Carl's use of stainless steel as a conductor for the comb.

But a new problem struck me as I was mulling over the legs. How to stabilise them, attach the drive motors to the wheel and still have full rotary motion. It seems that for greatest stability, the motor should be as close to the wheel as possible, allowing the shaft to be stabilised and supported at multiple points along the greatest length. This makes good sense, but sticks the motor way down in the dust and mud. Will this be a problem? I'm hoping that for the automotive-style motors I'd consider, the answer is no.

Thursday, January 7, 2010

To start with...

I thought I'd start with a simple description of what I want to build.

Start with 3 wheels, 3 drive motors, 3 turning motors. Place them in a roughly hexagonally shaped structure with the edges curved in an s. These "faces" would work to slide and lock together, allowing multiple robots ultimately to move around connected.

Drive motors would be attached to each "leg" to allow it 360 degrees of freedom. Slip/slide tines/rings could provide the conductive path to power the motors. Turning motors would sit in the main body and rotate the legs by chain or belt. Initial thoughts on locking mechanisms include automotive hood, trunk or door latch.


Major areas of concern include:
Comms mechanism: How to get the commands to the unit, and how to communicate these out to the necessary parts and pieces.

Drive: Translating drive commands to a coherent, cohesive set of commands to send to the motor controllers. This is a non-trivial challenge.

Sensors/input: This will become more important when we have more than one and/or have some sort of autonomous control, but knowing how far each wheel has traveled and exactly which direction it's pointing will be necessary for coordinating the forward movement.

Sensor/inputs I can think of currently:
- Compass (We can do a LOT with this) (absolute direction?)
- Accelerometer
- GPS
- Rough encoder for wheel rotation(?) - more connectivity, difficult/expensive pathways
- encoder for wheel direction.
- Camera

If we use a $100-200 eeepc as the brain, it has a built in camera, ethernet, wifi, bluetooth, USB, keyboard, screen, etc.

Makes me want to hack together multipoint bluetooth for the encoders on the wheels.

New Blog, New Purpose

So I have embarked upon this robot project. I've been talking about it for a year now and it's time to get something done with it. So my solution? More talk! Ya rly. I figure if I can come up with something to say about it every day until it's done, the project won't get stalled again. So look forward to parts musings, design twiddlings, and maybe even descriptions of actually building something.

The key is to make a post every day just to keep the momentum going. I'll have to be creative when I'm doing things like, oh, vacationing in the desert. Maybe it'll be done by then! (Or at least far enough done that I won't feel compelled to make a blog post every day...)