search

rouviecy / NAO_SLAM

Simultaneous Location And Mapping with NAO humanoïd robots
5 stars 1 forks source link

I am studying the NAO humanoid robot localization,could you tell me how to compile and run your program? #1

Open LiangLiang128 opened 7 years ago

LiangLiang128 commented 7 years ago

I am a college student, I am studying the NAO humanoid robot localization, find your project is very similar to mine, so I would like to learn the SLAM of NAO through your project.However, in the compilation and operation of the time there is an error, can you tell me how to compile and run your program? And what the dependencies required for the program to run (because of the opencv dependency issue during the run)?It would be so greteful for me if I can get your reply.

LiangLiang128 commented 7 years ago

Is there any related learning materials of your project？Because I wish to lrear about your project.Thanks a lot.

rouviecy commented 7 years ago

Hello,

It was also my project as a student, however I had not enough time to finished it. So right now, it does not process SLAM out of the box ; but only some experiments. You will find in this project tools to remote control your NAO (with keyboard or joystick) and tools to retrieve images from cameras, in order to process them on your computer (NAO CPU is not enough to do this). These are written in Python. My other repository "NAO_TOOLS" gives you remote control, and "NAO_SLAM/Visio/Python" gives you camera flux.

Image processing is another side of the project, written in C++ "NAO_SLAM/Visio/cpp". It is not compatible with NAO because a have not yet implemented the bridge between NAO/Python and PC/C++ (it's easy). So for this side, I usually used my webcam on my computer for the tests, or on another robot, not NAO. I was trying to recognize another NAO posture by vision, to do SLAM with floor tiles (not working with NAO, its head is not stabilized enough !), or follow things.

Oh, last thing : I used only vision, others sensors are cheating (part of the challenge !). I think now the best way to do SLAM with NAO is Hokuyo. But it was not part of my project. I will help you if you want vision SLAM, but I advise you rather use laser rangefinder : easier and more reliable.