Pololu Orangutan X2

Hi all,

I want to build a small robot, and wan to ask you if anyone have ever used the Orangutan x2 controller http://www.pololu.com/catalog/product/720 I see that lots of people are using the PIC controllers, but I don't wanna use these, they require windows and expensive compilers. I'd rather work with linux and program in C/C++

If anyone have used it before, is the orangutan easy to use? can i connect a camera to it without many troubles?

Also can you suggest a cheap camera that I could use with my robot?

 

Thanks a lot

Camera = tough

Interfacing a camera to any microcontroller is tough. About the only option is the AVRcam or one of the CMUcams, these do onboard processing and then output object data to a serial port. AVRcam and original CMUcam are each about $10l0. I believe the CMUcam2 is about $160, and the CMUCam3 runs upwards of $250.

The nice part about the AVRcam is that it is open-source and uses, you guessed it, an AVR microcontroller (note that there is a reason to put together an AVRcam rather than just slap the camera module on your orangutan, the camera interfacing eats up all of the processing of the AVR to do this, and has quite a few nice hacks and tricks to still run right…).

Only other issue I can see with the X2, is that it is an ATmega644, which I believe only has one serial port! Thus, if you connect an AVRcam, you will have to do software-serial for any X2-to-PC communications.

-Fergs

Some more cams

As above, the AVRcams and CMUcams have been used a bit in doing color tracking for a robot. They all make use of Omnivision camera sensors (OV6620 for AVRcam, CMUcam3, original CMUcam; OV7620 for CMUcam2)

Another possible is the Surveyor SRV-1 Blackfin Camera for $195, using an OV9665 camera read by a 500 MHz Analog Devices Blackfin processor. They have a stereo vision model too. I’ve had fun remotely running a BP Explorer robot setup that are Surveyor robots running around a diorama.

There are any number of wireless cams that could be used if that is all you are looking for, remote telepresence.

If you want to work with C

If you want to work with C and in Linux, the Arduino is probably a good option. Along with the Picaxe it’s one of the two microcontrollers frequently used by people on this site.

I don’t know anything about the Orangutan, but it seems kind of gimmicky to me. You’ll probably be paying too much for something that is underpowered and only meant for use with very specific projects. Not that the guys at Pololu are crooks, it’s just that it’s meant for a very specific use and for people without a lot of experience. If you go with something like an Arduino or Picaxe, you can learn it once and then apply your knowledge to all your future projects, while still being way more affordable and every bit as capable (and even more adaptable!) for your future projects.

You’ll also find more online resources for Arduino and Picaxe, making it easier to learn how to program and do cool things with.

… and welcome to the site!!!

Ooo that stereo vision one
Ooo that stereo vision one is shnazzy. I want one. Even though I would have no idea how to get Arduino to talk with it. Doesn’t matter, though. I want one. :stuck_out_tongue:

Thanks a lot for the info

Thanks a lot for the info everybody… i really like the stereo rig, but very expensive :frowning:

However, i’m a little lost… I’ll really appreciate it if you help me on this question

what parts you think i should buy to build a small robot with a camera? I need to know about those expensive parts, as it’ll be painful to budy stuff and never get it to work! motors, wheels and chassis are pretty easy to pick. My problem is with the controller and the cam

Thanks a lot

i love this forum, hopefully with some help with picking parts I’ll get my robot up and running

 

A camera is more for some

A camera is more for some more advanced robot builder. As a starter project I would rather use a infrared sensor or ultrasonic sensor. They are way less expensive and then the frustration is less if you can’t get it to work. But this will not happen as infrared and/or ultrasonic are very easy. Once you get the hang of that and learned the most important things about your microcontroller and how to interface with it etc. It’s time to try a camera :slight_smile:

GobliZ - Check out my robots

What to do?

What are you wanting to do with the robot, and why does it need a camera? What is the camera to do?

Are you just wanting to see what the robot sees? Or for the robot to analyze the video and operate from that?

Um, yeah, the X2 has a

Um, yeah, the X2 has a mega644 and a 168 (the chip on the arduino), it has hefty motor controllers built in as well. Thus, the X2 is actually about 3x the arduino in terms of I/O, processing power, and built ins.

The arduino would probably have more online resources, but the guys at Pololu have made a set of libraries for the X2 that rival the stock set of Arduino libraries.

-Fergs

Why even use an Arduino?

With those cameras, there is plenty of processing power left on the 500mHz blackfin! They have a set of libraries and tutorials on how to use it as your one and only processor.

-Fergs

As a builder who has a used

As a builder who has a used a camera with an AVR, I have this to say: don’t bother. Once you get to really using a camera, any of the solutions out there are probably going to frustrate you. The Surveyor model above, by itself, would be a good option. On the other hand, if you are really serious about doing vision with a robot (which is a really, really hard problem), then you might as well be prepared to pony up and just mount a webcam on a laptop and go from there…

Seriously, it took me about 2 months to get XRB3 up and running with an AVRcam. On the other hand, my latest bot is using a dell mini inspiron and a webcam, and it took all of 20 minutes to get it running - plus I can use open source vision libraries that far outweigh anything you find for smaller cameras. It’s just a matter of data throughput, you can’t move enough data through an AVR to be effective.

-Fergs

What I’m saying is that
What I’m saying is that that’s overkill. The Arduino or Picaxe would do everything he wants/needs to do.

the laptop idea sounds

the laptop idea sounds interesting … the problem though is that the laptop is gonna be big and won’t be able to get a small robot…

I’m aware of the intrinsic difficulites with vision algorithms, but I believe that the excitement and potential applications are stronger than the frustration!

I just want to small sized robot, no bigger than 1:10 RC car that i can put a camera on and navigate an unknown environment using vision only.

 

but that seems pretty complex!! what about a nono itx? I think that will get very expensive quickly!