Speech recognition - based control system for Drone

This project presents a speech-based control system for DRONE using Support Vector Machines (SVM). The set of controlling speeches consists of BACKWARD, FORWARD, HOLD ON, LANDING, MOVE UP, MOVE DOWN, TAKE OFF, TURN LEFT and TURN RIGHT are trained the SVM. The feature extraction of speech used in this study comprises of “fundamental frequency”, “Energy”, and Mel Frequency Cepstral Coefficient”. For performance evaluation, a set of features are used to test the SVM-based system developed by MATLAB. The results show that the average percentage of accuracy of the controlling speeches are 22.22, 46.67, 97.78 and 95.56 for fundamental frequency, energy, Mel frequency cepstral coefficient and all features, respectively. In addition, the interface of SVM-based system and DRONE is developed in practical use.