In this week's office hours, Sebastian said that he'd be more likely to teach a robotics course if there were really cheap robotics kits available. I was wondering what are the current options for cheap robotic kits that others are aware of?

asked 15 Dec '11, 19:18

melling's gravatar image

melling
1.2k211

I would love to take a robotics class using Hanson Robokind http://hansonrobokind.com/order/
(29 Dec '11, 22:30) dectal dectal's gravatar image
1
I have to wonder what kind of experiments can be done on the cheap? A $200 bluetooth, sonar enabled Arduino is a rather poor mapper without a better sensor (like laser range finder) or wheel encoders. But it would work. Maybe a remote webcam is the answer? Video or some kind of limited video sensor is a fairly powerful sensor. We need an open source hardware project to build the 'Stanford Trainer'? Would it be a mapper with a camera that processes on a PC? Would it recognize objects by imaging? What should it do?
(09 Jan '12, 22:14) esdad697 esdad697's gravatar image

15

Arduino? I think $100 gets you a decent kit. Toss on a few servos, and a few sensors.

link

answered 16 Dec '11, 00:09

mlepage's gravatar image

mlepage
1.0k8

Arduino is the most popular solution out there. And affordable. Hardware specs and software are both open. If one can afford laptop/pc and internet — he can afford Arduino. It costs less than some books.
(20 Dec '11, 09:02) vatula vatula's gravatar image
More like $200 with bluetooth so the processing can be done on a PC. And this does not include a laser rangefinder. It is difficult to do much on an Arduino. I made my own bot, but for the class a kit would be better (or prebuilt) so everyone could help each other. I have a couple of sets of Mindstorms too (for my son), but liked the Arduino better. The Arduino is all on a breadboard so you can modify it at will. Think open source hardware. Here is pic: https://plus.google.com/photos/105719296549854981266/albums/5691700909210536497/5691700909138868818 Treads cost about the same as wheels and are better on irregular surfaces. The entire bot is about 8" across which is very convenient. It uses the Ping! sonar sensor in the front.
(29 Dec '11, 18:50) esdad697 esdad697's gravatar image
I'd like to see something with Arduino. I'm trying to work with my daughters to build a couple of mini-sumos, one from solarbotics the other from parallax, both kits for about $100. The parallax one seems more flexible (programmable), but it's based on the Basic Stamp board and seems a little dated. Both of these have IR sensors, but would be nice to have more variety for same price. There is an organization called the First Lego League (http://www.firstlegoleague.org/mission/support) that works with kids 9-16 and is based on the Lego NXT and seems to work well for that age group, but it's all a bit pricey. I've just gotten into Arduino a few days ago, and the potential is mind-boggling. A $100 kit comparable to the 2 sumo kits mentioned above is probably feasible, but would be really cool to see what it would take to bring down a kit price down to $50 or less.
(01 Jan '12, 20:17) bjwiederholt bjwiederholt's gravatar image
I have one of the Arduino SparkFun kits which I purchased for around the $99 price point last year. It's more appropriate as an introduction to electronics. However there are various buttons, piezoelectric sensors and flexible sensors, so it could be a good introduction to sensors, and they do give you lights and a servo. You can also plug various other components in, including "shields", e.g. https://store.diydrones.com/product_p/br-0002-10.htm. Maybe check out http://diydrones.com/ to see what others have done.
(01 Jan '12, 23:37) aaronnewton aaronnewton's gravatar image

My personal vote would be for Mindstorms, but that's partly I already have access to a set that I gave to my son for his last birthday

Mindstorms has a lot of advantages 1) Very simple to assemble, robots can be mobile. Models can also be easily deconstructed with no damage to the set 2) There are a wide variety of sensors available, both supplied by Lego (touch, ultrasound, light intensity and color) and aftermarket. 3) There are a variety of outputs- the kit comes with 3 motors, a speaker and a small display, and there are other aftermarket ones 3) There are multiple possible languages you can run on the main brick 4) Bricks can be chained together to increase the number of sensors/actuators 5) Safe: no soldering or other tools needed.

It has two major downsides

1) It's not cheap in an absolute sense. A new kit pushes $300 and you'll want additional parts if you're doing complex things. You can get parts off of ebay, but I suspect the pool of available stuff will vanish quickly. 2) There are multiple versions out there that are only semi-compatible with each other

I'm not sure there's ever going to be an "affordable" kit in the sense of something sub $100 that includes a variety of sensors, actuators, brain and structural parts.

link

answered 16 Dec '11, 10:51

edremy's gravatar image

edremy
2506

How old is your son?
(19 Dec '11, 12:30) jimo jimo's gravatar image
1
He's ten. We got it for his birthday last year because our local school district runs a summer camp for gifted/talented kids and he'd been in a robotics group. Brag time- he and his partner built the only robot to successfully navigate the final maze. (I taught him the right hand rule :^) About half the class tried it by dead reckoning, which is absurdly brittle in the real world, and the rest used their sensors in odd ways. They didn't use Mindstorms, but rather BotBrain stuff: http://botbrain.com/index.html It was interesting, but honestly I think the mindstorms is better- the sensors are a bit more capable, the software environment seems a bit more stable and they are easier to assemble.
(19 Dec '11, 13:25) edremy edremy's gravatar image

Lego Mindstorms/NXT would be my first choice. A local robotics team uses The Finch for teaching programming but we haven't had as great experiences with the reliability of the sensors as I would have liked.

link

answered 15 Dec '11, 22:33

Jeanne%20Boyarsky's gravatar image

Jeanne Boyarsky
34117

yeah, Lego Mindstorms is probably best candidate - established, flexible, affordable, has bluetooth
(18 Dec '11, 23:55) EnTerr EnTerr's gravatar image
I think Mindstorms/NXT is a great choice. Simple programming interface, yet VERY powerful. Just do a search on Youtube to see the crazy things that "Legos" can do.
(20 Dec '11, 09:47) drm7 drm7's gravatar image

It depends on your definition of "cheap" and "good enough". There is the iRobot create (http://store.irobot.com/shop/index.jsp?categoryId=3311368) which is fairly neat. There are other robots available.

Almost any of these things can be made more awesome with an Android phone. If you get yourself and android phone and root it, you can use it to create a cellbot (http://www.cellbots.com/).

By strapping an Android to a Roomba or Create, you can build a powerful cellbot, which can drive around quite competently on flat surfaces (household floors, office buildings etc), and for a bonus vacuum your floors!

However, you're still looking at hundreds of dollars there ... say about $800 assuming a basic phone and a second-hand roomba. But you get quite a rolls-royce robot out of that. There are also bioloid robots available, plus Lego Mindstorms.

link

answered 15 Dec '11, 20:47

tleeuwenburg's gravatar image

tleeuwenburg
22014

$800 for a basic phone and second hand Roomba? I'd say more like $100 for a Create and $150 for a reasonably functional Android phone.
(16 Dec '11, 00:43) paxmaniac paxmaniac's gravatar image
2 questions about Android cellbot. 1. Could you do this from Android SDK emulator? (free?) 2. Do you really need to root an Android phone? The cellbots site mentions "RooTooth", but it sounds likesomething on the Roomba side.
(19 Dec '11, 12:02) genemiller genemiller's gravatar image
Not much point in using the Android SDK to drive a robot, IMO. May as well use the full power of your PC at that point. The advantage is really that the cell phone is a small, yet powerful computer you can easily attach to a robot. And it comes with a wifi and a bunch of sensors to boot. As for needing to root the phone: yes. It's the only way to get access to the USB port on the phone to communicate with the robot (through a microcontroller, I would assume), that I know of.
(19 Dec '11, 17:11) MikeDusseault MikeDusseault's gravatar image

Making physical robots work in the real world may be a lot of fun, but the hardware is expensive and ages fast. Plus there would be many limitations in what you can achieve with the hardware.

Simulation environments in software ('bots' with python/lisp/... brains) cost no money (that is, they may require a lot of work to set up) and can be almost as much fun, and maybe even more so, with competitions among students.

link

answered 17 Dec '11, 17:34

bm1's gravatar image

bm1
1296

1
Using a simulation is something to be considered, for sure. At the moment, anyways, I think it's the way to go. However, it should be cross-platform and run on at least Linux, Mac and Windows. Which doesn't seem to be a problem for the two environments I know (PyroRobotics and Nero).
(19 Dec '11, 17:15) MikeDusseault MikeDusseault's gravatar image

A robot definitelly could be built for around €20. And why not build a kit and sell it to the people who want to take part? What's the point of robotics when you even can not build a very simple robot?

Another option is a robot in the virtual world. An emulator.

link

answered 19 Dec '11, 10:57

kristjank's gravatar image

kristjank
798110

It seems to me there are several desirable characteristics. So far, none of the suggestions match all of them, assuming that the focus of the class would be on robotic software, not robotic hardware construction.

  1. It should not require a lot of skill to put together.
  2. There should be a good variety of sensors.
  3. It should be able to move.
  4. It should have a decent software environment. That means programmable in some normal language, running on a powerful enough processor with enough RAM that our programs do not have to be efficient. (Note that I'm not saying efficiency is unimportant in robotics! In general, though, efficiency and clarity are often at odds, and when learning you usually want clarity).

Of the suggestions so far, the TinkerForge stuff looks good on skill, sensors, and environment, but it looks like if you want a mobile robot, you have to build something yourself, using the module that can control motors. Perhaps some common motorized toy car could easily be adapted to work with it?

The Finch looks good on skill, movement, and environment, but the sensors seem limited.

Robotshop.com has a huge number of robots (complete and kits), components, and accessories. This Arduino-based system could be good. $90, and has a prototyping area available for adding a breadboard on which you could presumably add a variety of Arduino-compatible sensors.

They also have a bunch of chassis available, which are basically frames with wheels or tracks and the necessary gears and motors to drive them. Perhaps combining one of those, such as this one with the TinkerForge system would be good. Another promising base to build a robot on top of is this one sold by Sparkfun.

link

answered 16 Dec '11, 09:27

tzs's gravatar image

tzs
6786

(4) can be relaxed if robot/chassis has bluetooth or wifi connectivity. then the heavy thinking-lifting can be done by a smartphone (android/iphone/windows)
(18 Dec '11, 23:59) EnTerr EnTerr's gravatar image

Raspberry Pi is an option with an expansion board.

link

answered 16 Dec '11, 09:35

Govert's gravatar image

Govert
562

i don't see Raspberry Pi producing anything yet, so not ready for use
(18 Dec '11, 23:53) EnTerr EnTerr's gravatar image
1
Arduino or BeagleBone is better and available now
(19 Dec '11, 08:55) Merlin Merlin's gravatar image
Raspberry Pi will be available late January or early February and there is now a cheap board to interface with sensors and actuators which will likely be available at the same time much cheaper than Arduino: http://www.geek.com/articles/chips/raspberry-pis-gertboard-expansion-board-already-works-video-2012019/
(09 Jan '12, 12:30) AchilleTalon AchilleTalon's gravatar image

If Sebastian ran a robotics course that would presumably create quite a market for whatever kit was chosen ...

link

answered 16 Dec '11, 09:51

tim00d's gravatar image

tim00d
461

How about Roomba robot vacuum cleaners? They have motors and some sensors, and are apparently hackable via an "open" interface. In particular, the iRobot Create comes as a stripped down platform for creating robots using the basic parts (without the vacuum cleaner).

link

answered 07 Jan '12, 19:02

mlepage's gravatar image

mlepage
1.0k8

edited 07 Jan '12, 19:05

this is the most pragmatic solution! they have even released "iRobot Create" version without vacuum but with empty bay for hack-xtensions. So some can use the vacuum, some can buy special bot for the purpose
(07 Jan '12, 19:14) EnTerr EnTerr's gravatar image
Looks nice, the very basic version is $ 129.99 (without batteries) and $219.99 (with battery+charger). This looks like the cheapest/most accessible option right now.
(07 Jan '12, 19:35) Gundega Gundega's gravatar image

Merkur Robot comes to mind... Metal construction kits.

link

answered 16 Dec '11, 00:23

squared9's gravatar image

squared9
1964

TinkerForge looks pretty good. http://www.tinkerforge.com/

link

answered 16 Dec '11, 08:13

steve_corbett's gravatar image

steve_corbett
294

(22 Dec '11, 19:11) icogs icogs's gravatar image

Lego Mindstorms seems nice, plus it has good community and tutorials and not to mention lots of books.

link

answered 16 Dec '11, 09:31

calculemus1988's gravatar image

calculemus1988
2.1k1848

I think the best option is to create such a robot by hand and make it run lubyk with wifi. Such a robot could:

  • Run autonomous tasks (we only use wifi to debug, view sensor data, update code).
  • Do more complex stuff (we use the wifi to execute the calculations on a remote computer).
  • Have simple expansion slots (i2c, serial) to change sensors/actuators.

We could do everything by live coding Lua (a very nice and easy to learn scripting language).

If you think this could be cool, read this other thread: looking for contributors to lubyk.

For those inexperienced with robots, this setup has the advantage that there is no configuration needed apart from wifi password: no complicated development kits, no jtag, etc. Just a robot with code and sensors. We could make this an open hardware platform.

link

answered 16 Dec '11, 10:39

gaspard's gravatar image

gaspard
2.4k217

ugh... you clearly like this thing but this is not something on which to teach a class, it is something that has to be finished first. it's like saying i dont think we should choose programming language to practice AI, we should write a new one
(19 Dec '11, 00:13) EnTerr EnTerr's gravatar image
I've worked on something similar. You could remove the wifi and reduce costs further. use Rs232 for debugging.
(19 Dec '11, 01:36) iamsurya iamsurya's gravatar image
@EnTerr: I have been writing lubyk for 5 years now so I am totally biased.. ;-). I do not think Sebastian will teach a robotics class before Fall 2012 and this gives plenty of time to round the edges on lubyk. And by the way, Lua has been around since 1993 so I would not really call it a "new" language for AI. It's just the fastest scripting language around and it integrates easily with existing C/C++ libraries.
(19 Dec '11, 02:13) gaspard gaspard's gravatar image

Folks, what would you say about the following DIY hobby robot:

https://github.com/veter-team/veter/wiki/English-intro

https://github.com/veter-team/veter/wiki/Hardware-design-en

http://veter-project.blogspot.com/

I was building it with simplicity of hardware in mind, so it should be easy to make the copy. There is a lot of software already available but I am also keen to extend it using what I've learned in this class. Would be great if someone will be willing to join efforts on it.

link

answered 17 Dec '11, 17:20

Andrey's gravatar image

Andrey
1818

Thumbs up for the idea and work done on this. Sadly it does not fit into the range of cheap kits.
(07 Jan '12, 19:41) Gundega Gundega's gravatar image
Thanks for the positive response! Regarding price - you are right, it is not cheap, but there are much more possibilities with this platform. In contrast, the really cheap platforms (like those based on micro-controllers) quickly become boring after you implement line following and simple obstacle avoidance (as already done by millions of hobby developers). More powerful platform is required to implement really interesting algorithms and unfortunately it costs more money.
(08 Jan '12, 10:51) Andrey Andrey's gravatar image

He could also teach the concepts without specifying a hardware platform.

Suppose that the course requirements state that the student has to be able to control up to 3 motors using a computer, and sense 6 digital inputs and 2 analog inputs. Also, the course requires a stable carriage (3 wheels or more) with drive capability, and an unstable carriage (2 wheels, balanced) later on.

We could leave the platform up to the student. The student can use an Arduino, but they can also use a TI or a mindstorm or even a laptop. The carriage could be the size of a mouse, or a roomba, or even an old car.

The environment could be a "maze" made of common items chosen by the student. A "standard environment" could be a square 30x the size of the robot on a side. Then the course could specify requirements within the environment without worrying about the construction details.

For example, you could place 4 2x4 boards on your living room, and then place shoeboxes for obstacles. You could tape together 8.5x11 sheets of paper together to make a line-following robot.

Homework could be as simple as a video of the solution in action. I would really enjoy seeing the creative solutions people come up with.

link

answered 18 Dec '11, 17:31

Barrabas's gravatar image

Barrabas
6.5k828

I think I will focus (the right word...) on computer vision for a while. The reason is that I can concentrate on software writing, without any work on the hardware. A single webcam is enough to begin. I have succeeded today in installing opencv on ubuntu 10.04 and 11.10 on two boxes. If someone needs some clues, you can ask.

My Trust webcam is recognized, so now I must read quite a bit about the working of opencv. I installed also the python support, so I think I will use python for the first tries, then I will switch to C to have a good performance.

Since my system is multicore, can someone suggest what to do to use effectively all this processing power? I head about the TBB libraries, but I haven't tried them yet.

I am also thinking about a small mobile robot, and for this I think the best thing is building it from scratch. A possibile approach could be an Arduino as a local interface to motors and sensors and a wifi connection to the main software on a remore pc. Or a small netbook installed directly on the robot. But of course there are plenty of possibilities.

link

answered 18 Dec '11, 17:56

Finix's gravatar image

Finix
1.1k217

edited 18 Dec '11, 18:28

Arduino seems a good basis, and they have relatively inexpensive bundle kits (around £100), but every time (at least 3) I've tried to install the driver software on Windows, it's seized up my system. It is very popular at hack days for prototyping robots (phones on roller skates etc).

I feel the lego mindstorms are a bit limited in scope, already having a DSL etc. Not sure you could get much more than 2-3 lectures out of it. Any on-line robotics course would need to use parts that are widely available.

What about robot simulator software instead?

link

answered 18 Dec '11, 18:43

NaN's gravatar image

NaN
3655

Here is another interesting project http://romotive.com/

link

answered 18 Dec '11, 22:25

krokodil's gravatar image

krokodil
3247

It requires a smart phone but $90 for the mechanics to move the robot seems like a good price. I was thinking about getting one just to experiment with the AI I learned in class. I'm not sure on how programable the robot is.
(10 Jan '12, 18:14) jhager jhager's gravatar image

I've worked on simple robots from scratch and can build them in about 20$

They're 8051(NXP) or Atmega based, have 4 motors + sensors and can do the following:

Obstacle detection, Line following, Sumo robotics.

There's more you can do with it once you get the hang of it. Can be used with Microsoft's Robotics Studio.

If you're looking for cheap, the best thing to do is import cheap parts from China / India and then let everyone start from scratch. If you're looking for easy then something based on the Arduino should do it.

Media: Similar "robot": http://www.youtube.com/watch?v=iPKkAF99kLo

Our first board: https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-snc1/10035_148023845107_535185107_3230353_1462001_n.jpg

The final board looked much better. No, really ;)

link

answered 19 Dec '11, 01:50

iamsurya's gravatar image

iamsurya
605

I'm just buying old toy robots and modifying them ... bolt on a few parts here or there ... job done

link

answered 19 Dec '11, 08:56

Merlin's gravatar image

Merlin
4.2k51429

It would be good to know what are the minimum capabilities of the robot that Sebastian requires in order to teach a class. We are talking about a lot of work for Sebastian to prepare a class. Maybe he or someone could get a manufacturer to give deep discounts to his "students". The chances of this happening anytime soon is if Sebastain is already familiar with the robot. I have some hope because Sebastain seems like a guy who likes to do things and not just talk about them.

link

answered 19 Dec '11, 15:00

jimo's gravatar image

jimo
558213

2
Just what I was thinking Jimo. Get students to sign up an interest, then go to a manufacterer and say 'How much for 10,000 of these kits?' That should bring the cost right down, and then we'd all be working on exactly the same kit.
(19 Dec '11, 15:32) dowsabella dowsabella's gravatar image

Cheap robotics? Break out the Legos! Seriously, I know a kid who has one of those robotics kits and they are rudimentary, but with some serious modding could be made to work with whatever language is chosen to program the 'bots.

link

answered 30 Dec '11, 14:55

Tasega's gravatar image

Tasega
797

Would love to get links to anyone's code. Going to hack on AI classwork during break between sessions. This session coding was focused on ML class. Anyone wanting to collaborate can find me on Google+ -- Dan Scott of MindTree (lots of Dan Scotts)

link

answered 16 Dec '11, 10:13

danscottphx's gravatar image

danscottphx
594

Parallax (http://www.parallax.com) has different affordable robotic platforms for hobbyists in the $100-1000 range that (at least in United States) can be found in brick&mortar stores (microcenter, frys) in addition to online.

For example Scribbler, Boe-bot, Sumo-bot and Stingray are withing $100-150.

On the higher end, I remember i saw at Silicon Valley Code Camp a Microsoft Robotics platform that included base, Kinect sensor and a laptop - i think it was this thing: Eddie Robot ($1250)

link

answered 19 Dec '11, 01:03

EnTerr's gravatar image

EnTerr
956

edited 19 Dec '11, 01:04

in Russia it's Robopica. You can write programs in MicroC for it. It doesn't cost too much (when I bought it 3 years ago it cost about 5000 roubles - that's something about 156$).

http://www.terraelectronica.ru/catalog_info.php?ID=1171&CODE=246178&Name=IE-ROBOPICA&Razdel=%CD%E0%E1%EE%F0%FB%20%F0%EE%E1%EE%F2%EE%F2%E5%F5%ED%E8%EA%E8&TableName=class_19_5_1

link

answered 19 Dec '11, 03:21

indra-uolles's gravatar image

indra-uolles
314

For most of the robotics class you do not need the robot limbs, but just a mobile wheeled platform with some IR distance sensors and a CPU board. The whole thing can be assembled from parts which are sold already in places like the one here, less than for 200$.

My own interest is to go further and use my Android phone with a USB peripherials to control actuators/read sensors. I think of using the USB accessory library and a relevant arduino board for this. Also WiFi and camera on the phone will come in handy when it comes to robot vision.

For those who are a little lazy, I would suggest buying a Rovio mobile webcam. It has a web/API which allows to control it, get an image, etc. You can control it from a PC and do robot vision tasks.

link

answered 19 Dec '11, 06:47

petkish's gravatar image

petkish
7267

I assume you've seen ADK. If not, take a look. Alternatively, you can attach BT to Arduino or buy the board with the one on it.
(19 Dec '11, 16:02) Pacman comma... Pacman%20commander's gravatar image
link

answered 19 Dec '11, 12:55

Pacman%20commander's gravatar image

Pacman comma...
843210

http://www.phidgets.com/

More flexible, but possibly more expensive as well. Not familiar with Mindstorms, but I do know the Phidgets have built-in USB capability and have a well-documented API. They've also been available for at least a couple years.

Oh, and the robotics class has my vote. Where do we sign up? :)

link

answered 19 Dec '11, 16:27

VCCGeek's gravatar image

VCCGeek
865

Sebastian may not be doing a course on robotics, but the Stanford Engineering have this one available: http://see.stanford.edu/see/courseinfo.aspx?coll=86cc8662-f6e4-43c3-a1be-b30d1d179743

link

answered 22 Dec '11, 06:25

NaN's gravatar image

NaN
3655

Thanks, @NaN. "... basics of modeling, design, planning, and control of robot systems. ... results from geometry, kinematics, statics, dynamics, and control. ... Prerequisites: matrix algebra."
(30 Dec '11, 03:23) EllenL EllenL's gravatar image

There is also the option of having one or more robots available at Stanford for the course, that could run the software written for it by the students (online or at Stanford). Maybe after passing a more or less stringent (depending on capacity) selection test for viability, to make the best use of the limited resources at Stanford for the Robotics class.

Because, if you build software for vehicles (say, Volkswagen), you're not the one who builds the vehicles, or even the required tooling for measurement and control, but your software is only deployed on vehicles. Same with cameras, smartphones, etc.

So, a combination of initially simulation, and then real deployment (with collection of video and data of course).

/not suggesting Sebastian should 'donate' one of the cars of course, as it would probably be crashed, maybe with collateral damage :-) But toy cars, maybe. Software for collecting real-time data in well known data structures and for operating motors etc could even be pre-delivered, to keep down to realistic proportions the amount of work necessary to achieve interesting results and lively competition among participants.

link

answered 19 Dec '11, 11:37

bm1's gravatar image

bm1
1296

Because, if you build software for vehicles (say, Volkswagen), you're not the one who builds the vehicles, or even the required tooling for measurement and control, but your software is only deployed on vehicles. Same with cameras, smartphones, etc. Actually I just watched http://www.youtube.com/v/uoiJeIb0wBA (The great Robot Race), a link posted at the reddit forum about this course (thanks for the link!) and learned that Sebastian and his team exactly did this: concentrate on software as much as possible.
(19 Dec '11, 12:53) bm1 bm1's gravatar image

A lot depends on whether the "robotics course" is going to really deal with the mechanical engineering aspects of robot design. If not - if the course will be mostly concerned with software, then LEGO NEXT would be a really good choice.

OTOH, if the course will address the mechanical aspects, then the LEGO solution may not be a good option. It really restricts your mechanical design options, especially if your goal is a robot with a fairly small footprint.

FWIW: After the lectures on particle filters, I Googled and found a few particle filter implementations for the NXT. Haven't had a chance to try one out yet, though...

link

answered 19 Dec '11, 12:21

philko's gravatar image

philko
66517

Raspberry Pi have ARM based thingy running Debian Linux for £15 Uk or $24 - mad specs.

They just announced a £15 add on for controlling tons of servo motors and lots of inputs for sensors and the like.

http://www.raspberrypi.org/archives/411

$50 for a compete low powered debian with loads of servo/ sensor ready I/O

link

answered 20 Dec '11, 08:39

Jason_42's gravatar image

Jason_42
3477

Let's wait till it definitelly moves out of vaporware status.
(20 Dec '11, 09:23) kristjank kristjank's gravatar image

Raspberry Pi is certainly something to watch for the future, though in its basic state it lacks sensors and actuators--just to get feedback from a running program you'll need to attach a HDMI monitor or an analogue (RF) television.

Gert van Loo is working on a general purpose I/O board for this machine, but it isn't anywhere near production at the moment and there are no concrete plans to sell assembled versions.

http://www.raspberrypi.org/archives/411

Another possibility is Arduino. This is much lower-tech than Raspberry Pi, but it can directly drive actuators and sensors.

http://www.adafruit.com/products/50

link

answered 20 Dec '11, 10:22

Tony%20Sidaway's gravatar image

Tony Sidaway
3385

Raspberry Pi is supposed to be available for sale at the end of the month or early next month. Also, there is an expansion board for sensors and actuators which is likely to be available at the same time from another source and much cheaper than Arduino: http://www.geek.com/articles/chips/raspberry-pis-gertboard-expansion-board-already-works-video-2012019/
(09 Jan '12, 12:32) AchilleTalon AchilleTalon's gravatar image

My vote also for Raspberry Pi. It is, in fact, not yet in production, - almost, but not yet - as it is (will be) a very nice piece of hardware, with several interfaces, and capable of running Linux (the few prototypes have done a good amount of success at the Qt Developer Days. It has many quite important features for a development device: it possible to have a decent GUI (it has a OpenGL capable GPU) several connectivity ports (ethernet, USB, serial), and as one of the most important: it is unbrickable ;-)

link

answered 22 Dec '11, 07:14

frares's gravatar image

frares
3056

Isn't Lego Mindstorm progamming choices quite limited?

There is Arduino analog, but with 32-bit, 80 MHz chip (approx 100 times faster than original 8-bit, 20 MHZ Arduino). Price is comparable to Arduino.

http://www.digilentinc.com/Products/Detail.cfm?NavPath=2,892,893&Prod=CHIPKIT-UNO32

This should be more suitable for serious projects

link

answered 22 Dec '11, 18:20

Alar's gravatar image

Alar
103

How about using Wiimotes for cheap infrared cameras, for both robot vision, and generally as a sensor (e.g. wall detector or floor line finder). They are readily available and hack-able.

link

answered 01 Jan '12, 23:48

mlepage's gravatar image

mlepage
1.0k8

Raspberry Pi has gone to manufacturing

Availability expected at the end of January 2012.

link

answered 11 Jan '12, 12:23

AchilleTalon's gravatar image

AchilleTalon
1.7k11

Your answer
toggle preview

Follow this Question via Email

Once you sign in you will be able to subscribe for any updates here

Q&A Editor Basics

  • to upload an image into your question or answer hit
  • to create bulleted or numbered lists hit or
  • to add a title or header hit
  • to section your text hit
  • to make a link clickable, surround it with <a> and </a> (for example, <a>www.google.com</a>)
  • basic HTML tags are also supported (for those who know a bit of HTML)
  • To insert an EQUATION you can use LaTeX. (backslash \ has to be escaped, so in your LaTeX code you have to replace \ with \\). You can see more examples and info here

powered by OSQA