A trained machine learning model made using Edge Impulse and downloaded to work on an Arduino Nano BLE33 Sense. Used to turn on a light by using someone's voice. Works by being added as a library and then being accessed using an example with access to the microphone.

To use the trained model access and compile an example using either a microphone or continuous microphone. When it is running speak into the microphone and it will attempt to classify what is being said. Use the serial monitor to find out the results. They keywords it can detect are - 'Light On', 'Light Off', 'Livingroom On', 'Livingroom Off', 'Bedroom On', 'Bedroom Off', 'Kitchen On', 'Kitchen Off'. The solution is not perfect and will occasionaly mistake words for others such as mistaking 'Livingroom On' for 'Bedroom On'.