My Mobot for the 2010 Mobot race has finally been completed. I have tested it electrically, so I know that part works (all of the sensors!). I’m working on writing code for the race (just over a week away!). I’ll add more pictures and talk about progress as it’s made. Hopefully I’ll get some video of it in action!
Stairbot is getting closer and closer to completion. The above render is what the final version should look like. We’ve decided to laser cut a lot of acrylic to build most of the leg and body structure.
We’ve ordered all of our electronics, and we should be receiving them soon. Then we can start actually programming and getting gaits set up. Our goal is to have the construction mostly completed by the end of the month, and start working on some basic motion control.
Below are some images from the build process.
I’m in a class called Introduction to Robotics. In this class we have labs where we make robots to do things using the LEGO NXT Mindstorms kits. (Part of) Our current assignment is to make a robot that can follow a fairly complex line. In our team, I’m the Mechanical Engineer, so I built the robot. Our Computer Scientist, Rich, wrote a really sweet line following program based on the general algorithm our group came up with.
I’m not going to reveal the black magics he used in his program, or our general algorithm, but, from the video you can probably deduce the general idea. The line it follows is complex, containing both a 90 degree bend and a hairpin. As you can also see in the video, the robot could follow this line indefinitely until its batteries die.
I managed to get a gallery of a step-by-step build of the MicroColony/Build18 sensor board uploaded. This post details the sensor board itself a bit better.
We plan on building a better one that uses a multiplexor to allow us to have more than 5 sensors on the board.
During Build 18 I managed to slap together a simply chassis for our micro colony robots. It was super simple, made of some plastic and hot glue, but, it was able to hold everything on it.
Here are some pictures:
As you can see it isn’t very big, not much larger than my cell phone. I used 3 4-40 machine screws which let us slide the sensor board into place and tuck all of the electronics underneath. I really like the 4-40 screws for stacking feature and it is something I would like to retain for the next revision of our chassis.
The next revision will also have a mount for the servos, instead of them just being tacked in place with hot glue. It was a quick and dirty build, but, it got the job done.
I’ve managed to complete 1 sensor board for the build 18 project. The basic idea of the sensor board its to let the robots figure out their relative positions. The sensor board consists of 5 IR emitter/detector pairs. The basic idea is one robot will turn on its IR emitters and the other robots will use their dtectors to figure out their relative position based on which detector registers seeing the IR.
The sensor board is based, in concept, on the one found here: http://ipvs.informatik.uni-stuttgart.de/BV/swarmrobot/tikiwiki-1.9.2/tiki-index.php?page=sbII . However, our sensor board uses only 5 IR emitter/detectors and we do not use ours for communication, only for relative positioning.
We plan to have at least 2 of these working, hopefully 4 in total.
Here is a gallery of images from the construction of the board:
[nggallery id=6]
Building robots is expensive. Motors $20, controller $20, motor driver $20, sensors $10, battery $10. Just the basic components and you’re already up to $80. The costs add up fast and it’s hard to pour a lot of money into building a lot of robots. That’s one reason why it’s great to have a resource like the Robotics Club. If I wanted, I could walk in and build a complete robot with parts that are already there. All I would have to invest is time, no money (well, a $25 membership fee, but that’s beside point).
Because building robots is expensive, for now I’m going to be focusing on projects that don’t require funding or that I have funding for already. I’m going to focus on Mobot (which I have almost all of the parts for), Micro-Colony (which I’m getting funded through the Build18 project), StairBot (a robotics club project I’m the head of, I’ll keep updated on its status as I work on it), and some other projects that I have in mind.
Unfortunately, I’m pushing Desk Rover to the side for the time being. Mainly because my robot fund is a little short of the price tag of desk rover, and I have plenty of other robots to work on in the meantime. I will build the first stage (chassis, sensors, motors, etc. no LCD or wireless) once I have enough money.
I will continue to work on the FWR3 Jasper. Most of the work in it is design work as of now, and that doesn’t cost anything but time (which, I now have a lot of as I’m on break).
I am planning on participating in Carnegie Mellon University’s Build18 event (http://build18.org/). The week long event is hosted by the Department of Electrical and Computer Engineering. The idea behind the event is to promote electronics tinkering. I’m going to be working together with Derek Kozel (http://www.derekkozel.com/).
Derek and I are going to attempt to build a colony of micro-robots. These robots will be approximately 2in x 2in x 2in. We plan to build at least 4 of them. These robots will be able to communicate with both wireless and infrared beams.
These robots will also be equipped with solar cells which will allow them to recharge themselves. Our goal behind this micro-colony is to have the robots work together to find the brightest spot, the location where they will be able to recharge most effectively.
I’ve wanted to build a colony of small robots for some time, after having read about this one: http://swarmrobot.org/. I’m really looking forward to the even and the chance to build this colony.
Today I found out that CMU’s RoboClub’s Colony robots have encoders which could be used for the Jasper platform. The colony robots use the same motors and wheels that Jasper will use. The encoders are magnetic (not optical). This will save me a lot of development.
I still need to do a lot of development with the board. But, the robotics club might have PCB fab supplies which would speed things up greatly. I need to talk to some club members to see what we have.
Over the past few days I’ve been thinking about completely redesigning the Fun With Robots platform. New chassis, new board, everything. I’m going to redesign it to make it more versatile. Giving it more features and greater capabilities.
I’m calling this new platform “Jasper” and designing it around the old platform plus an additional feature set. Jasper will have similar features and functionality to the FWR 2 platform, which is based on the Firefly board. In addition to those base features, the Jasper platform will have expandable capabilities. It will be able to use add-on boards similar to the Arduino Shields. The first big add-on I’m looking to add is optional encoders and then maybe wireless and LCD add-ons.
After that the possibilities are endless. I would also like to rework the curriculum some and maybe market it to other colleges and maybe even highschools. All in all it will be nice to get a new platform out, and it would be great if we could get other institutions to use it. Expect to hear more about it in the near future.