Cyberflex
Cyberflex
CAD has been printed for just the hand, while the CAD for the wrist and stand has been worked on.
We recognized a few problem with the hand that will be fixed with new CAD.
We have learned about optimal muscle placements for the sensors. We have also written code for the Arduino which can now move 5 servo motors connected to the Arduino to individually control fingers and differentiate between rock, paper, and scissor gestures. We also created code to use our trained model to predict the gestures, but we haven't been able to fully implement it yet due to issues with our data collection.
We have created machine learning models which can successfully differentiate between different gestures to a high degree of accuracy. We also began work on a script that can detect these gestures in real time. We hope to be able to implement these models in our next sprint.
To create a new 3D printed mechanical arm with a forearm block that can act as a stand. We also hope to change the fingers and motors to DC motors based on what the mechanical team decided to use in the CAD model. We will also fix our script to read sEMG data, predict a gestures using our model (which we will also train on more data and number of gestures), and translate it to the motors to control the hand. We will also update sensor placements for each of the electrodes to optimize muscle signal readings.
This site was created with the Nicepage