SDDEC20-18 • Magic Door Sensors

Create a system that can detect, with a reasonable amount of certainty, the state of the doors in an environment.
Battery powered security devices have built in maintenance and wired devices can be expensive to install. Our goal is to use the waveforms of certain wireless signals to determine if the propagation in the environment has been effected by any changes.
We used a machine learning model known as Random Forest to analyze the packets and detect the likeness of different angles.
We also implemented a secondary, interrupt driven, electronic switch that will be used to double check and continually train the model. This system will be powered by a supercapacitor and a passive collection antenna.