Skip to content
No description, website, or topics provided.
C++
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
combineintovector.unit
nn.unit
nn
obstaclesprofile.unit
opencvredobject.unit
writetrainingset.unit
README.md
gather_training_data.rvw2
nno.rvw2

README.md

General information

This project was developed during the Bachelor thesis at the Hochschule Mannheim. Main idea is to navigate the robot "Robotino" produced by Festo Didactic in an unknown environment.It should follow an object with the given properties and try to avoid the obstacles on its way. Instead of using directly the Robotino API custom function blocks for the program "Robotino View" have been developed.

Image processing is being done using the OpenCV library. The controlling of the robot is accomplished with a simple feed forward neural network using the FANN (Fast Artificial Neural Netowrk) library.

Links to the documentation

Here is a short guide how to get the openCV library work in Windows and make it accessable in Visual Studio 2010
Install and use OpenCV

Here is a short guide how to get the FANN library work in Windows and make it accessable in Visual Studio 2010
Install and use FANN

Here one can find the information about the data sets, test sets and network training Train and test a neural network

Todo List

  • FB NN. Setting of the path to the configuration file through the FB gui.

  • FB WriteTrainingSet. Setting of the path to the training set through the FB gui.

  • FB CombineIntoVector. Fix the bug with the last output being 0 if the last input is empty.

You can’t perform that action at this time.