3d game studio and Real Robots

There is someone thinking in the idea of work with 3dgs for programming real robots?

it is possible?

Dude I don't think that is possible so far. 3d game estudio is game simulation enviroment and so far as I know to program a real robot you need to understand real low-level scripting code oriented to events and SED is oriented to objects. At least that's what I know.

i wouldnt call it totally impossible since you can access the serial device and thus can controll something external

I don't think it is possible - it just sounds too unrealistic...

I build robots myself and program them with "C".

Why would you use 3DGS anyway? It's power is in the 3D rendering IMO, and this isnt needed to program a robot. Standard C with a good IO library is probably the way to go.

The only sense in using a 3d engine for this purpose is to simulate behaviour in 3D and adapt it to real robots (like puppets) or as visualization tool for realtime action or as viewer for logged robot session to analyze misbehaviour or something like else.

If you do this, keep in mind that you have to built 3D models and animation skeletons similar to your robot and that you have to deal with euler angles and so on (gimbal lock!) when translating rotation matrices to gamestudio entities.

Best regards,
-Christian

Thanks, Christan and friends. I think that my question is not a "crasy" question, because 3dgs is based and compatible with C languaje and really when we program 3D characters, vectors and behaviors we are programming virtual robots.

The second reason is because CONITEC makes products like GALEP series, a device that uses to program industrial chips and so on.

Thank , friends.

If you want to work on real robots, get Microsoft Robotics SDK in a language you think you can work well with. My personal choice is always C# or C++. I've never made a REAL, PHYSICAL robot that exists in the real world, but I've made virtual bots before in C#.

I'm actually, in my free time, playing with making a "chatter box" bot which will talk to the user and holds conversations with them. I'm using Microsoft's Speech Recognition and TTS/SAPI Dlls (System.Speech and its children namespaces). It currently recognizes a couple hundred phrases (and interprets whether they are questions/greetings/commands/etc) and responds to you (by voice & text) with a pretty realistic decision making branch.

The raw C# programming is close to finished, it's just the massive library of "knowledge" that will take ages to develop completely. I'm using a new Markup Language I created in XML, which is similar to AIML (<- research that). This makes it so I don't have to "hard code" string after string. Even non-programmers can "teach" the bot things by adding new markup (very easy). When I have time, I'm going to add some new features, like dynamic learning. When the bot doesn't recognize what you say, it will ask you to "teach" it some proper responses; which you then dictate to it, and it adds to its markup library. That will eliminate the need for any more programming AT ALL. You just "teach it" like a child.

This isn't as hard as it might sound at first (if you're moderately experienced in programming). I suggest that if you want to start making robots, you should get the MSR SDK and learn C# (you can get it all FREE). Try some simple AI bots first before you spend thousands on parts kits. You'll have to ask experienced eletricians/robotics engineers & hobbyists about making the actual robots from parts (I know very little about that, except how to fire events from the robot sensors to your code). But once you learn the basics of AI and get familiar with your language, you will learn FAST. I wish I had more time to play with bots, but real life and "WOP" (work-oriented programming, lol) suck away the time.

If you decide to follow my advice, you can always ask questions! But just go ahead and toss the idea of Lite-C/A7 for robotics. Sure, you could FORCE it to work, but you're going to need to write new C++ libraries (Which defeats the purpose of Lite-C). Plus, WHY? It's a game engine, not a robotics SDK. If you wanted your robot to display video or be controlled by a virtual 3D interface, then shoot for good ol' DirectX (or even XNA + DX). There's a lot you can do, but you don't need any game engines; you need good libraries and a strong IDE (like Visual Studio or Visual <blank> Express). Java is also an option, even though I don't care for it. You can use practically ANY language that doesn't "shelter" you from low-level development by a super-high level of abstraction. Though it's a beautiful and elegant scripting language, Lite-C can't cut it! *cough* C#! *cough*