Bachelor Theses - Details
Adaptive Active Noise Cancellation for Headphones using Virtual Sensing
Betreuer: Jonas Förster
Themengebiet: Signalverarbeitung, Active Noise Cancellation, Adaptive Filterung
Schlagwörter: Active Noise Control, Virtual Sensing, Adaptive Filtering
Tools: Matlab, Simulink
The field of active noise control (ANC) aims in reducing or even cancelling unwanted noise at specific
points or even areas in space. Especially ANC headphones became very popular in the last decade as
their performance increased noticeably due to recent advances of digital signal processors (DSPs). Nevertheless, mainly time-invariant filter designs are still used in modern commercial headphones.
The use of adaptive algorithms allows for even better performance as they are able to adapt to variations of the headphone fit and the direction of arrival of the incoming noise. Thereby, the signal energy at the error microphone, which is usually implemented close to the headphone speaker, can be minimized. However, the actual goal is to reduce the signal energy at the users ear instead. While the placement of the error microphone is limited to the headphone, virtual sensing can be used to minimize the signal energy at a virtual sensor placed on the ear drum.
In this thesis, virtual sensing shall be applied on the headphone-specific case of an adaptive ANC
algorithm. For this purpose, the necessary transfer functions from the error microphone to the
inner-ear microphone of a head and torso simulator have to be determined by acoustic measurements.
Based on these paths, an adaptive ANC algorithm shall be extended by virtual sensing using Simulink.
Eventually, the performance of the algorithm shall be evaluated compared to conventional adaptive