The Design And Implementation Of Active Noise Control In Vehicle Cabin

Noise, vibration, and harshness(NVH) is an important factor in the development of modern vehicles. One of the challenges is the control of low frequency noise inside cabin. Active noise control(ANC) is a promising solution since it is good at reducing low frequency noise. However, the method will experience some difficulties such as low convergence rate and not working well for broadband noise control and so on when applying to vehicles noise control. This is a key factor to restrict the development of ANC inside vehicles cabin. In order to improve the performance of ANC system in dealing with the vehicles’ broadband noise, this dissertation is mainly concerned with the design and implementation of the control algorithm.Based on the acoustic wave linear equations and the superposition principle, the single channel feedforward system are used to explain the principle and structure of ANC system. The filtered-x least mean square(FXLMS) algorithm is derived and the possible factors which limit the performance of the traditional FXLMS algorithm are analyzed from the views of convergence rate and steady-state misalignment. Meantime, the effects of the performance indicators of the control algorithm on the noise control are analyzed. In combination with the data collect and analysis for the vehicles’ interior noise, numerical simulation for the sound field inside vehicles is conducted using finite element software LMS virtual LAB. The noise levels, frequency spectrum, sound filed distribution and so on are analyzed in the space around the right ear of driver under different work situation. Based on the impulse response of the secondary path is sparse, the relationship of sparse and proportion factor and convergence rate is analysed. this thesis introduced the idea of proportion into the traditional modeling algorithms., the precision and convergence speed of the modeling algorithm is improved.Based on the relationship between correlation of the input signals and the performance of control algorithm are analysed theoretically, with the characteristics of noise inside vehicles cabin, two improvement methods: variable step-size parameter design and decorrelate the input signal are presented. With reference to the Leaky-FXLMS algorithm, a control algorithm is presented. In this algorithm, variable step-size parameter is introduced in the cost function and the step-size is optimized in the process of determining optimal weight vector. In this context, the performance of noise immunity is improved. Meanwhile, the subband adaptive filtering is considered in the control algorithm. The input noise signal is partitioned and decimated via a filter bank to reduce the correlation of the subband input noise signals. Thus, the convergence rate of the control algorithm is increased. Moreover, this thesis uses a recursive formula to reduce the computational complexity in order to facilitate the realization on the microprocessor. Simulation experiments are performed using narrowband noise and broadband noise as input signal. The results show the subband adaptive fitering technique is effective for broadband noise control.A one-dimensional sound field based on Digtal Signal Processor development platform is set up. The verification of the secondary path modeling algorithm and control algorithm is completed with the application of narrowband noise and broadband noise. Experiments based on a single channel ANC system are implemented inside the vehicle. It is seen the improved control algorithm may effectively control the noise.