This package contains a set of nodes that are used to get user handwriting input and adapt the next word to the skills of the user
Tested with ROS Indigo on Ubuntu 16.04 (LTS)
$ cd ~/catkin_ws/src/
$ git clone https://github.com/chili-epfl/choose_adaptive_words.git
$ cd ../
$ catkin_make
See CoWriter project requirements
In ./nodes
run python activity.py
to launch the writing app.
The main node that opens two windows, one for writing the other one to manage the parameters
- Change the Simple Learning Pace (0,100)
- TODO : use PCA
- TODO : use CNN
- TODO: use adapt thibolt
The launch argument choose_adaptive_words
For instance, for a specific experiment, this way:
$ roslaunch choose_adaptive_words [...other options]
- Remove Federer Button (be able to change to adaptive mode or regular mode with a threshold we could set
- Clean screen just after the we send a new word
- Make a "real" eraser, select the zones to clean
- makenao look at tablet
- check arms
- make the path of the robot editable by the kid