Frequency domain inverse model LMS algorithm for active sound tuning system of powertrain noise

Traditional vehicle active noise control (ANC) systems aim mainly at attenuating the interior noise and usually do not account for sound quality effects. Active sound tuning (AST) system is proposed to reshape the vehicle powertrain response based on the predetermined vehicle interior sound quality criteria. Since the computational cost in AST system is important in real-world application, the frequency-domain filtered-x lease mean square (FDFXLMS) algorithm is utilized. However, the conventional AST system cannot yield a balanced performance over broad frequency range and good convergence due to the inherent frequencydependent convergence behavior caused by the dynamics of the secondary path. Therefore, a novel AST system configured with frequency-domain inverse model LMS (FDIMLMS) algorithm, which utilizes the inverse model of the secondary path to eliminate its effect on the algorithm convergence, is proposed as the basis for tuning vehicle powertrain noise. Compared with the traditional AST system, the proposed system not only yields faster convergence speed and more balanced performance, but also reduces the computational cost. To validate the effectiveness of the proposed system by tuning individual engine order response, numerical simulations using measured powertrain noise signals were performed under both constant and run-up engine operating conditions. Results show obvious improvements in convergence rate and balanced reduction over a broad engine rotational speed range.