This document proposes an audio-based speech recognition system for use in vehicles using a KNN classification method. The system aims to reduce driver distraction by allowing drivers to control common vehicle functions like temperature, windows, phone calls, and GPS through voice commands. It uses Linear Predicted Coefficients (LPC) and Mel Frequency Cepstral Coefficients (MFCC) to extract features from training data consisting of driver names and commands. A KNN classifier is trained on these features and tested, achieving 80-90% accuracy for driver identification and 50-96.6% for command classification. The system shows potential for increasing vehicle security and reducing accidents caused by distracted driving.