Investigation On Multi-channel Active Noise Control Algorithms For Near-field Error Sensing

There are two basic ways to reduce noise.There are passive noise control(PNC)and active noise control(ANC).PNC deals with th high-frequency noise efficiently.However,the efficiency of PNC drops down when the noise frequency is low,due to the high cost and difficulty in deployment.ANC by using microphones,and loudspeakers to generate the low-frequency anti-noise has become a complement to PNC.Besides the success in the active noise cancelling headphone,a novel application targeting the urban traffic noise is one of the focusing topics in ANC,namely the ANC window.The ANC window is able to maintain a quiet living environment while keeps the window open for natural air ventilation.The noise passing through the open window is regarded as the primary noise source.The error microphones can be placed in the farfield,which is beyond half the wavelength of the noise from the open window.In this case,the noise power is approximated by the sum of squared errors and classic FxLMS algorithm is optimized to minimize the noise power.However,the far-field placement of the error microphones occupied the interior space of the room.Apparently,The ANC window prefers the near-field placement of the error microphones.Without the particle velocity measured in the near-field,it is critical to solve the problem how to get the equivalent noise reduction performance using the near-field pressure sensing as compared to the far-field pressure sensing.In this thsis,the fundamentals of the ANC algorithms are firstly investigated by simulating different affecting factors.Secondly,the multi-channel virtual sensing method is proposed to solve the a formentioned critical problem of the ANC window.Thirdly,the minimax method is adopted to reduce the computation complexity of the multi-channel ANC algorithm as well as the multi-channel virtual sensing method.Last but not least,an experimental platform is built up for the validation purpose.