Laptop or microprocessor

Hi,

I'd ask what you want the robot to do...If you want it to be autonomous I'd go for a microcontroller.

You cam have the best of both worlds by using a microcontroller which can be re-programmed via a development board linked to your laptop. There are plenty of microcontrollers which can be re-programmed in 'C' which will give good flexibility.

I don't think you can beat the fun of having a robot run completely 'hands off' anything with an umbilical dngling off looks like cheating...but hey that's just me!

Del

It is autonomous, based on a Q learning algorithm.

The robot's purpose is to avoid obstacles and find an optimal path to travel and reach its destination.

Hi,

I would agree with Catman, FYI, I work with 6802 designing robot which is running at 8 speeds in forward and reversed direction way back in the 80's. The project required me to designed the controller and the driver board as well. It was tedious but full of fun and you feel satisfied when the project completed.

With microprocessor your robot would be free to roam, which I believe more practical.

However, if the project required massive programming power then I guess you got to go with the Laptop.

Good day to you,

I had come to conclusion your robot has to carry its "brains" either laptop or controller itself. Otherwise your post should be "PC or microprocessor". It would look very funny robot holds laptop at one its hand and by means of other one keying-in commands! .

But to be more serious.

(1) I do agree with Del the Cat and other participants absolutely as I would prefer to use C for any like projects as well. C gives us both relatively human-friendly environment and very smart and compact object code. Though there are a good brands supplied FBD programs which are believed to be quite suitable.

(2) I'm assuming you have some experience with both LabView and Matlab. As for LabView, I do not like it, especially for control applications. For scientific instrument measurement --- yes it would be good but for robot?..I have a bit of doubt. But it's only my personal preference.

(3) Matlab. Yes --- it's look pretty workable. Though Matlab does deliver practically endless set of approaches how to execute your project. I suppose you will go as most of users to compose some Matlab lang code and Simulink. If you are skilled at Stateflow --- it's very good and it saves you a lot of time. Then (I bet) you're going to generate PC(Windows) target code for compiling Matlab real-time obj module which presumed to communicate with bot through RS232 or Ethernet. Yes? But in any case you must have here some "iron" device which can accept data from COM port or Ethernet. Isn't it better to apply (1)?

(4) If you are going to use Matlab OPC function (SCADA protocol). I would advice you to shun this approach, though I do not think it's your real intention.

(5) I could suggest one more one win-win decision. Not me but Mathwork. Check Matlab list of supported devices. Amongst them you can find just controllers (not embedded boards). If it affordable for you. Buy one of them and you can download there Matlab generated C++ code thought to be the same compact and smart. Therefore on early stages of your project you can use comfortable environment of Matlab installed on your PC (not a laptop). Then you end up having downloaded debuged c-code to controller and be enjoyied of your bot's good working.

But I remind you again. You may be casually sudden to get unexplainable problems and bugs having to use these high integrated programs. So see again (1).

In case (1) nobody gives you assurance you will create workable code, but everyone assures you that always it's possible to find any bug and fix it.

I wish you success at all beginnings.

Thanks a lot..!