This paper presents the experimental validation of a wearable driver motion monitoring system that uses Inertial Measurement Units (IMUs) placed on the head, neck, and torso. The system measures linear accelerations and angular velocities to automatically detect the driver’s driving behavior. Controlled experiments were conducted in a traffic-free urban environment. Participants performed a continuous driving task that included common maneuvers (straight-line driving, braking, rounding, lane changing, yielding), simulating four different driving styles: normal, aggressive, distracted, and drowsy. A Recurrent Neural Network (RNN) model with Long Short-Term Memory (LSTM) units was implemented for behavior classification and demonstrated high discriminative capacity between driving styles, achieving 99.33% accuracy. These results validate the effectiveness of the proposed system for driver monitoring and accurate detection of driver behavior.