For the adaptive active headrest system of road noise in a vehicle, the secondary source cannot generate sufficiently large sound in the low frequency range, which may lead to excessive amplitude of the control filter causing distortion of the secondary signal, and then affects the system performance. For the feedback system, the performance is also affected by the noise amplification caused by the waterbed effect. A sensitivity constrained filter error LMS (ScFeLMS) adaptive algorithm is proposed in this paper, which can constrain the amplitude of the sensitivity function of the feedback active noise control (ANC) system, thereby suppressing the noise amplification caused by the waterbed effect. The combination of the FeLMS algorithm also suppresses the low frequency output of the controller. Using the transfer functions and noise data measured inside an electric vehicle which is driven on rough road at 50 km/h, the noise reduction performance of the adaptive feedback system and the adaptive hybrid system based on the proposed algorithm is simulated. The results show that the control performance in the target frequency band from 70 Hz to 500 Hz of the adaptive feedback system is similar to that of the multi-channel feedforward system, but the system complexity and cost are significantly reduced, while the hybrid system can significantly improve the noise reduction in the target frequency range.

Adaptive algorithm; Feedback system; Hybrid system; Electric vehicle; Road noise control