The document describes a real-time voice actuation system that uses a source-filter model of speech and support vector machine classification to recognize vocal commands with high accuracy in real time. The system uses a TMS320C5515 DSP chip to handle vocal input and feature extraction via autocorrelation, transmitting data over Bluetooth to a Raspberry Pi for classification using LIBSVM and actuation via GPIO pins. The finished system demonstrates functional voice recognition of the commands "one", "two", "three", and "four".