LMR, global project, possible ?
2008-05-02 08:46
Hi everybody, I just had this crazy idea, and where else if not here to post it.
What about if we all build the first online open source robot as a community? I’ve seen some projects and profiles of the members here, and there are professionals, students and enthusiastic, which means that we all can contribute in the project.
I explain myself better: The objective of the project would be to create a robot where everybody contribute, sending Programs, Source Code, Electronics Schemes and real applications for the robot to create the first robot (as much complete in applications) made 100% online by people from all over the world in the LMR community. This would be advertising for the LMR site and who knows, perhaps we all become famous and we would be part of history because we would create the first robot as a global and online community.
The goal would be to present the robot in some international congress like the ‘’Robocup’’ or the ‘’Eurobot’’ by next year (2009) and also publish the advance with some white papers and articles. Also we would meet each other finally on that event.
I guess that we all can find sponsors to afford the construction of the robot and the main idea would be to keep the robot free and in an open source way to let people know our work and keep the contributions for the future.
I know this sounds crazy but if is not here, where else can I post my idea?
Just think about it.
Daniel
- idaniel's blog
- Login or register to post comments
I think that open source is
I think that open source is the future, and community work is the best way to get it on, yes, we can think in something like that, something that can run in a lot of difrente machins, like a Robot Vitrual Machine (comapring with Java VM) that can work with the hardware that we have, so each one of us can make our own test robot at home and contribute, and we can also make a big machine to use as final project to all the parts we done.
Have i expressed me well? think that ppl that know about java will understand it!
Just comment so we can brain-storm the ideas and go on.
Absolutely!
Where do we start? I se the following tasks (In no particular order):
1) mechanical design
2) electrical design
3) software design
First, though, we need a consensus on what the basic functions should be including drive (wheels, tracks, legs, etc.). Perhaps we should decide on a weight or other bulk? Let's do SOMETHING...
Perhaps anyone who's interested might post a list of their abilities? A kind of Curriculum Vitae.
I'm a software engineer. I can program PICs in their native RISC assembly. Not so hot on C, but I do program embedded control computers for a living. Also, an electronics enthusiast and I dabble in mechanical engineering.
God created the integers; all else is the work of man - Leopold Kronecker
Maybe we create a set of
Maybe we create a set of modules which are pluggable? Maybe based on a common bus? Say we pick I2C, for example. We'd build I2C interfaces to an RS-232 converter, a servo controller, an ultrasonic sensor, a radio control receiver, a 20A motor controller.
That would be soooo sweet.
God created the integers; all else is the work of man - Leopold Kronecker
i agreed with the I2C, think
Robot Task
I like the I2C idea and the robotic modules. It would give us some standard guide.
I guess we should start thinking on the applications that we would like to develop for the robot.
Here some ideas:
Line Follow, Obstacle Avoid, Gripper, Multiprogramming Task, Internet Protocol, RF Control, Space Recognition, Reprogrammable Robot, Learning Task.
Please give more ideas, then we pick some applications and we start thinking on how to build the robot.
PS: I like the Curriculum Vitae idea, I’m Mechatronics Engineer. I can develop the entire virtual robot so we can give us an idea of how it must looks like. I also can help on some mechanical advices for the design and some C++ programming stuff.
Daniel
Well, we can think in maping
Well, we can think in maping system, using walk + compass + sonar sensors to keep making the zone, maybe in future with video, and image recognition. and possibly with gps module but don't think is important, since its 2 easy to work with, nothing to explore, and is limited to outspaces.. i prefere the indepentend maping!!!
I'm Informatic Engenhiering student, i love electronics and software..
follow a person, guide a
Currently...
"Follow a person" is my current project anyway (http://letsmakerobots.com/node/655). I'd be happy to share the platform. At the end of the day, I would hope that a "modular" solution should lend itself to anything. For example, in my current project, I'm using old windscreen wiper motors from cars. I would be happy to make the driver board for that an I2C device, so that if we develop some shared code, somone could plug in a driverboard for a pair of smaller motors or a pair of stepper motors or the like.
Ultimately, if I write code to control it, someone should be able to plug my code into their 'bot and it should work. The core controller could be identical to mine (Microchip PIC based), but since we're going modular with a common bus, someone else might choose to use an Atmel board and rewrite the software for that: the I2C motors drivers would be pluggable.
God created the integers; all else is the work of man - Leopold Kronecker
What about a topo mapping
Early Days
I'm not sure I would give it a function yet. For my money, I think we should prepare a list of functions which could be added. I would hope we could describe a "Sensor and Control Array" based on I2C, handing out responsibility to various groups for the design and documentation of each element. Once we know what inputs and outputs we CAN manufacture, then most of the hard work (and probably most of the coding) is already done. All that remains is to script some high level code around it.
...and this "high level" code should be done in a fashion that a partially-trained monkey could use it.
I ony suggested a person follower because that's what I'm working on and I'd be happy to manufacture it from generic modules.
I DO like the idea of a LARGE mapper, though. The next logical step from tha being an electronic postman! YES!
God created the integers; all else is the work of man - Leopold Kronecker
I may not have the electrical part.
Computer Science student at
Computer Science student at University of Technology, Eindhoven. I'd love to help with the programming.
"Lead, Follow, or run around like crazy"
My only question is how to
My only question is how to best collaborate (and listen Ice is back with my brand new invention.... sorry!!!). Do we work together on the web? Or does 1 person make something and mail it to the next that adds somethign else? Do we make a $$$ limit on how much someone can add? I think it would be fun to get a box in the mail, open it up, (save all mailing labels so we know where it went) add something, and mail it to the next person. Eventually we would have a robot with a well traveled passport and possibly something cool.
Also it would be cool if everyone had to make 2 improvements: 1 functional (add a motor, some code to do something, etc) and 1 cosmetic (add paint, glue something, etc personalize that adds a bit of everyone who did something)
i think you didn't
This is also what I
This is also what I envisage, although, I REALLY do like idea where a box of bits goes around the world!!!
God created the integers; all else is the work of man - Leopold Kronecker
I actually love this idea. I
I actually love this idea. I know it's not exactly what the original thread was about, but I'd love to work on a collaborative robot like that, especially with the two stipulations you mentioned in the last paragraph. Of course, then only one person ends up with the 'finished' robot. I still think the 'christmas robot exchange' idea sounds like a great idea too, then everybody ends up with a robot they get to keep :) Maybe combining the two... Organize a 'ring' of people, and everybody starts one robot and mails it to the next person on the 'ring', and when a robot gets back to the person who started it, they get to keep the robot. Or maybe the person that keeps the robot is the person just BEFORE the one who started the robot -- the person who would otherwise mail the robot back to the first person. That way the robot you keep is a robot you didn't work on at all, so it'd be entirely novel and foreign and interesting. Either way, I'd LOVE to be involved in that project :)
Dan
An Opensource Robot
THIS is sexy: http://oap.sourceforge.net/.
$2,000 is pretty big bucks, though.
God created the integers; all else is the work of man - Leopold Kronecker
Good ideas
Hey, I like the ideas. In one hand the objective would be to build everything via ‘’web’’, means we all put our talent, so the robot will be like ‘’free open robot’’, so any of us can build it by their own, I mean someone has to build the robot with all the specifications and the stuff and can upload the process of construction. It means that the mechanical design (planes and measures), the electronics schemes (board and connection components) and the programming (source code) will be open so it would be just to add into the robot. I guess depending on the personal free time and the economic facilities the idea would be that everyone builds the same robot, because it will be a very complete robot where all of us would contribute so it will be a robot with a lot of applications.
In the other hand the idea of one robot travelling around the world sounds really cool. Our robot would travel more than ‘’ASIMO’’.
Now I guess we should start thinking on the robot itself, a robot to do what?
Daniel
NOT what to do...
The "to do what" question is one I've been trying to avoid. In fact, earlier I alluded to figuring out what "modules" we are capable of building and putting the end result together later. I would be tempted, for example, to build an I2C motor controller aimed at the 12V 20A range. The controller has no purpose other than to be included in a robot which wants 12V 10A motors.
Where to draw the line is one question. There are already I2C ultrasonic distance measuring devices and compasses, etc. I feel , however, that they are extortionately over-priced and believe that a $40 ultrasonic sensor could be manufactured for less than $20.
In many cases, I suspect the software to be the most expensive component of these systems. I once wrote a program for a friend who offered to pay me for the work. I calculated that three evenings a week for 2 months at an average engineering rate of $60/hour, the program would cost him around $3,000. He nearly choked, but because I enjoyed writing the program, I settled for a nice bottle of wine instead.
God created the integers; all else is the work of man - Leopold Kronecker
i really think that the I2C
i really think that the I2C would be the best way, so imagin i wanna make a walking robot instead of a driving robot, i only need to change the motor driver module and use a servo controller module, and imagin that after that i wanna add him a extra well to be able to drive, its just to plug another module that have already been made by us, and adjust the software to use it.
Yes, its true that most of the sensores are overprised because of the size, and the engenhiering process. We can together make cheaper sensores, so we can use much more sensores, that can inclusively be better ones, we can use betters ADC, or ways of talk, like in a utrarage we can set up I2C to ask for "what ditance to a object?" and sensore modlue reply the value, or we can ask "is anything in your midle rage?" and he reply a true or false. We can basicly inprove this a lot, and have cheaper modules that we can build ourselfs, and understand whats going in the down level, but add better funcionaltys.
If we work well as a big team, we can do whatever we can imagin.
I go strart thinking seriosly at this.. strating with the low level microcontroller programation. My dear turtle bot will need to wait ^_^
World Time
By the way, I was just thinking and also would be really cool the Time Zone Difference. While in one part of the world someone makes the contribution on a day, the others are sleeping and when they wake up the next day, the robot will have something new to work on, while the others are sleeping and waiting the new contributions...
Daniel.
So, who’s in ?
Who is in?
How do we start?
Daniel
I'm in
Me.
To start, I have two devices on the drawing board.
1) I intend to develop an I2C 2-channel 12V 20A reversible motor controller. The aim is that it be suitable for driving two car windscreen wiper motors ineither direction with a range of speeds (more accurately "torques") and offer some limited speed feedback using the wiper motor's own inbuilt "once-per-revolution" switch.
I am undecided as to wether to provide scope for this to be extended for better rotational position or speed feedback, or to develop that as a seperate device. My inclination is to develop it as one single I2C module such that it could be given the instruction: "rotate forward at 40RPM" and the main CPU would not have to bother about it.
2) I am a big fan of using as high a level as possible for the main processor and it is my intention to enable the use of a PDA as the "brains" of future projects. I have already done this with great sucess using a Palm III (which you can get on eBay for $60). It should be a reasonably simple excercise to separate my protocol (which I will freely share) from an existing project and put it on a Microchip PIC (favoured MCU) which natively supports RS-232 AND MSSP. (MSSP because this device would be the I2C bus master.)
I have also used a Psion 3c and a 5 for past projects and their inbuilt programming languale (OPL) is brilliant and dead easy to learn. I settled on the Palm III purely based on its small size. is a great device for this. I have used yBASIC AKA HotPaw (free download: http://www.hotpaw.com/rhn/hotpaw/) with great success to address the serial port. Of course the project should be extensible to allow interfacing with more modern PDAs via USB.
Ideally, I would use a Palm Vx, or something else which could be powered from the main batttery of the robot, but I'm on a budget.
God created the integers; all else is the work of man - Leopold Kronecker
Additional
If I might add a third item, I have developed my own 16-channel servo controller, which currently has a serial interface. It should be straightforward to convert this to an I2C device as well. I will happily share it. You won't like it, though: like all my projects the MCU is coded in RISC assembler.
The other thing, re the above message: If this takes off, and anyone else is interested in adopting the idea of using a PDA to control their projects, it would be great if you could all lobby the guy who wrote THIS version (http://www.aldweb.com/articles.php?lng=en&pg=5350) of BASIC. It is definitely superior to yBASIC and the IDE is excellent, BUT it has no serial port support!!
God created the integers; all else is the work of man - Leopold Kronecker
i'm in, i'm starting with
i'm in, i'm starting with PIC to be able to make I2C modules, so i will share all the info i will learn with this, and i think we can make nice modules, for sensores, and dc or steper motors controllers.
When i have more info i tell you guys, i'm just going to buy some PIC and other stuffs to be able to program them, and strart on the PICs world.
keep this project going, with everyone doing him best, i can't expect nothing else that a greath final result!!!
Don't Spend Money!!
Ahhh. Noooo! Don't BUY PICs!! Register at eh Microchip website and request free engineering samples. (They even pay the postage!) Select your PIC carefully from the BRILLIANT on-line interface. If you want a slave I2C device, get a PIC that natively supports the interface. Anyway, Microchip will send you up to 3 of any of 3 different products in one shipment. With 9 PICs for free, and a free IDE you can't complain. (Except you'll have to learn RISC assembler to be able to employ the IDE.)
God created the integers; all else is the work of man - Leopold Kronecker
thank for the warning, ok,
Hi, an open-source robot is
I think it would be fun to