Prediction of Interior Vehicle Noise by Means of NARX Neural Networks

In recent years, great interest on NVH characteristics of vehicles has been paid by all the big automotive manufacturers. Interior acoustic comfort is now one of the main key factors in vehicle development process, since it contributes to improved product overall quality. Therefore, in automotive industry advanced NVH refinement needs to work in synergy with all research activities. Assessing the level of experienced noise in interior cabin requires particular arrangements for ensuring adequate measurement accuracy (AC system off, closed window, etc.). The use of parameters such as the level of seat vibration, not affected by the acoustic field conditions inside the vehicle, could facilitate experiments in parallel with engine/vehicle calibration activities. These parameters, in addition to information about engine/vehicle speed conditions, may be used to indirectly determine the interior acoustic level in real time, by means of a prediction model based on limited acoustic measurements inside vehicle cabin, previously performed. Present work describes the development of such a prediction model, properly tuned on the basis of the noise and vibration experimental data acquired inside a passenger car cabin, tested over a track in both stationary and transient operating conditions. More in detail, a Nonlinear Autoregressive with External Input (NARX) Neural Network was implemented for the prediction of the interior noise level over time. The good overall performance as well as the possibility to generalize new data proved the remarkable prediction capabilities of the trained network in a real-world forecasting scenario.