We prevent Roboy from falling over. By mimicking the patella tendon reflex, Roboy is able to detect a critical and ustable position and stand up to a more stable one.
Currently the critical values are recorded from a highly unstable knee position. This can be extended by providing more unstable positions and implementing machine learning, kinect vision detection, etc. to optimally calculate the critical position. Also a IMU sensor placed at the hip can help with stabilizing the hip upward motion.
Furthermore, an Intel Edison should be used to attach the whole functionality to Roboy. A whole PC setup to start the reflex is thereby avoided.
Our template is run with ROS catkin and can be extended to different reflexes.
##Prerequisites:
roboy_ros_control roboy_ros_trajectory
for recording data: roboy_gui
##Building
To build the project you need to clone the project into the catkin workspace src folder
Example:
cd ~/catkin_ws/src
git clone https://github.com/valentindiehl/Roboy_Reflexes.git
cd ..
catkin_make