• Multi-Array Switch add-on, consisting of 4 magnets that can be moved contactles into two different states. 16 different states in total are therefore possible. Magnets are included.

• The additional needed 3D2Go board or the 3D shield can be used for the read-out.

With a magnetic 3D sensor a contactless multi-array switch can be generated. In this example, four independent moving magnets create unique magnetic fields indicating 16 different switching combinations. This can be used as a Human-Machine-Interface and has no tear or wear to the sensing element.

The multiswitch code contains the following files:

1. data.txt
2. dataCollection.py
3. main.py

pip install {lib}

• numpy
• pyserial
• xgboost
• bitstring
• pandas
• sklearn

This program uses a machine learning algorithm to predict the exact multiswitch array combination. Data needs to be collected and then feeded to the XGBOOST algorithm to train on it. We provide a ready training dataset in data.txt. Run dataCollection.py and follow the instructions if you want to create your own dataset or extend the already existing one.

Finally, run main.py to predict the array combination. When trained on the provided data, an accuracy of 95% (+- 3%) is achieved. You can play with the algorithm's hyperparameters and feed more data to it to further improve the accuracy.