Rather challenging as I am still learning C. Lots of learning ahead, and that’s the big attractiion afterall. 
I’m learning about the “switch, case” in C. Very interesting and soon implimented. I just need a little time to read and digest.
Thanks Pam.
Well here it is with an updated chassis and a servo to sweep the PING sensor left/right.
I don’t have the servo operational yet but the sensor and everything else is online. I’d like to try an idea I have to add bigger wheel directly to the existing wheel hub. That will be the next phase of development.
I’m spending the day here at my work bench figuring out the pin outs on the Roomba main board. There are 18 connectors and 85 pins so it’s going to take some time.
I’m finding some very interesting things here. For instance, there’s a micro switch on the front wheel that is pulsed on and off as the wheel rotates! The Roomba I’m working on was bought on Ebay as a unit with “problems”. As it turns out, the front wheel on this one is seized up with hair and crud and this causes the drive motors to behave erratically so that’s probably why this one ended up in the junk pile.
My main goal here is to identify the connectors for the IR sensors located on the front and then determine what voltages are present and how that voltage varies when the sensors are triggered. Then I’ll be able to utilize these “cliff” sensors on the 4 wheeled Roverbot to aid in autonomous navigation.
When I’m done I’ll be posting all this data to my website to share with other robots builders that wish to hack the Roomba hardware.
I’ve been working on this prototype rover robot which proved to be a worth while endeavour.
I found the motors in the Roomba to be very user friendly. The have a planetary gear and belt driven assembly inside that provides very good torque and they can be driven directly from the H-bridge that is already on the Roomba main board!
I mounted four of these on the suspension arms, two on each side. They were then connected to the mainboard at the factory plug. I did however load up the circuit with double the normal load because I was using two motors per side. It seems to handle this just fine.
The Arduino microcontroller I installed sends a PWM signal to the trigger transistor on the H-bridge of the main board which is powered by the Roomba battery pack. Mounted to the front of the chassis is a Parallax PING sensor which tells the controller when there is an obstacle.
Here’s a video of the prototype in action
http://www.youtube.com/watch?v=xrPBWtjbHKs
I’m currently working on a more final design to take to the 2010 Maker Faire in San Mateo in May. I’ll be building a new chassis and suspension assembly out of aluminum and making use of one of the IR edge sensors from the Roomba. I’m working on a really neat wheel driven by the stock motors… stay tuned.
I’ve installed some gear reduction motors I had laying around. I used some 9/32″ brass tubing to slip over the output shaft of the gear box, then I drilled a hole in the shaft and installed a set screw. After drilling the wheel hub hole out to 9/32″, I inserted the axle and motor assembly through the drilled out bushing in the suspension arm. Next I pressed the wheels onto the axles by hand. I soldered a twisted pair of wires to each motor and routed them up to the batteries on the main chassis.
It crawls over things fairly well considering the lack of torque and over all weakness of the gear box.
So, I now have proof of concept for a six wheeled drive train. Now I need to find some higher torque gear motors.
Here’s a video of it in action:
http://www.youtube.com/watch?v=FxE4sayLlV8
I found a light weight fold up baby stroller at a yard sale for $7.00!!
The wheels that came on it are perfect for this prototype. I’ve mounted them up and they fit great. The tire is made of high density closed cell foam so they should grip well enough.
Now it’s all about fabricating the drive mechanism. I might experiment with some gear reduced electric motors I have just to test the traction. Eventually, I want stepper motors on all six wheels, preferably built into the hub. We’ll see about that one… so far my searches haven’t turned up anything affordable. Electronics “salvage yards” will be the place I’ll end up finding what I want.
I had a very busy build session today involving a trip to the hardware store for some more hardware. I found I had less play in everything if I used a bolt with an unthreaded shank on it. After retrofitting a few, I built the box chassis with a cross member for the rocker arm. This is a component that lends itself to the name Rocker – Bogie Suspension. I mounted the suspension arms and rocker and gave it a wiggle. It worked great!
Next, on with the wheels. For now, all I need is a simple wooden wheel, so, I re purposed a painted pine plank I had laying around and cut out six four inch diameter wheels. After a little sanding, I installed them on the legs and gave it good look. I put it in a few different configurations in an attempt to simulate what it might go through in a boulder strewn terrain. It performs quite well and deals with a wide range of positions thanks to the way the rocker links the two suspension arm assemblies.
Of course, when you start looking at a prototype, things pop up you didn’t expect. That’s part of the process though so I pay attention in those moments even though my builder ego may be bruised. Ego schmego… work the problem. Anyway, you start to think “what if a drive motor failed on one of the two bogie wheels? What keeps it from just snagging on something and rotating 360 degrees and ending up useless?? ” Yeah.. what if? So now I’m thinking a limit of travel is in order, but I would need some data on maximum angles and such, hence, this prototype.
So now I can collect suspension angle data , operating parameters, and continue forward to the next phase, powered locomotion.
I took my PVC pieces into the machine shop and drilled all my 1/2″ holes for the nylon bushing to press into. They’re all straight and fit tight!
The trick here was to keep everything on plane while assembling the PVC with the solvent weld cement. There’s very little time to work so planning ahead makes a big difference. I pre assembled and adjusted everything, then I put a line on each joint with a red Sharpie marker. I rigged up a jig to help hold everything in the right place then set about gluing it up!
I’m very pleased with the results. I have the right and left suspension arms complete for less than $25. These are prototyype arms and they won’t steer, but they’ll demonstrate proof of concept.
Next task is to attach wheels and then bolt both arms to a central chassis. After that, there’s the rocker arm to install that links the two suspension arms.
PVC pipe is easy to obtain and it’s inexpensive. So, I’m building the prototype suspension parts with 1/2″ PVC pipe and various fittings.
I drilled holes in the pipe at the pivot points and inserted some 1/2″ x 1/4″ nylon bushings along with some 14″ x 20T bolts.
Each side took about 20″ of 1/2″ PVC, 4 - 90 degree elbows, 3 - 45 degree elbows, 2 nylon bushings, 7 washers, 2 bolts and 2 nuts.
Next I’ll add the arms that extend down to the axles and the wheels. Then, the whole assembly gets mounted to the main chassis.
Click the pics to enlarge them.
All Terrain Robot = ATR.
Well, the Maker Faire is over and I proudly walked away with an Editor’s Choice Award from Make Magazine for my Rumble Robots with an Arduino. I built two more while there and had a real blast with everyone. Now it’s time for a new project!
I’ve set a new robotics goal for next year:
Build a roving robot based on the six wheeled Mars rovers. There are three generations of these rovers. The latest is called the Mars Science Laboratory , AKA “Curiosity”, and it’s big! About three times the size of Spirit or Opportunity. All three generations of the Mars rovers have a common component. The Rocker Bogie Suspension. I really like this suspension design and I’d like to experiment with building one so here we go!
I’ve started by producing some drawings:
Most of the mechanical construction of the chassis involves the suspension components. The main body is just a box that houses the electronics and batteries and carries a solar panel on it’s back. After a few more drawings are completed, I’ll begin building a prototype model to test the suspension geometries.
I’m still considering materials for the chassis. Aluminum, steel, PVC pipe, composites… all have come to mind. Steel would make welding things together very easy but I would have it’s weight to contend with. Plastic and composites are a lot of work and expensive. Aluminum would be great but I don’t have easy access to a TIG welder. I like the idea of PVC because it’s so readily available. For now, I’ll just focus on building a scaled down prototype.