Active Noise Control of Airborne Road Noise Based on Artificial Intelligent Computed Road Classification and Adaptive Digital Filter.

The study presents a new method for real time noise control based on road surface classification technology using deep learning and active noise cancellation technology using digital adaptive filter. Recently active noise control (ANC) method for structure vibration induced road noise is developed and applied to real car for road noise control. This method is effective for ANC of road noise due to structural vibration of car body but is not effective for ANC of road noise due to air borne noise source such as interaction noise between tire and road surface. Acceleration signals measured on the car body have been used for ANC of structure borne noise. However, a new reference signal is required for ANC of air borne noise. Interaction noise between tire and road surface is severe on the concrete road than the other road such as asphalt. The reference signal should be correlated to interaction noise between tire and road surface. In the paper, the type of road surface is classified in real time throughout deep learning of the measured interaction noise based on convolutional neural network (CNN). The reference signal of interaction noise on concrete road was obtained by calculating instantaneous frequency of interaction noise and was successfully applied to ANC of airborne induced road noise.