Acoustic applications of high-performance adaptive filters for FPGA platforms
Publication date
2025-10-07
Document type
Dissertation
Author
Advisor
Referee
Granting institution
Helmut-Schmidt-Universität/Universität der Bundeswehr Hamburg
Exam date
2025-05-14
Organisational unit
Publisher
Universitätsbibliothek der HSU/UniBw H
Part of the university bibliography
✅
Language
English
Keyword
ANC
Speech enhancement
Active impedance control
FPGA
MIMO
FxLMS
Helicopter headset
Abstract
One of the great challenges in modern society is the reduction of noise and vibration. High noise levels can occur in urban environments or industrial applications. Whether the goal is to protect humans from the negative effects of noise, enable communication in extreme environments, or acoustically disguise vehicles, traditional passive dampening is not always the best option. The concept of active noise and vibration control has been in existence for several decades, yet it has only been adopted on a widespread basis in certain applications, such as headphones. When larger acoustic volumes need to be controlled, active control systems with many components must be applied. In these systems, the computational effort of the algorithms, coupled with low latency requirements, presents a significant challenge. Hardware-based active noise control systems can be a solution due to their high level of parallel computing and low latency. In other areas of signal processing, they have already become the industry standard for the same reasons.
In this thesis, acoustic applications of adaptive filters implemented on Field Programmable Gate Arrays will be presented and evaluated. The developed systems in this work include well-known setups, such as Kundt’s Tube or an active headrest, for validation purposes. The more advanced solutions incorporate a helicopter headset with additional speech enhancement capabilities and active impedance control of an underwater surface. Additionally, an overview of the theoretical background is given, and state-of-the-art related work is discussed.
The results of the measurements show that hardware-based adaptive filters not only match their software counterparts, but they can enhance noise reduction performance. In the case of the helicopter headset, the additional computational headroom of the chosen platform is used to include Wiener filter-based speech enhancement. This system produces impressive results in the evaluated speech quality, in addition to the active noise reduction of the headset. Active control of underwater sound bears additional challenges when it comes to latency requirements due to the high speed of sound. The proposed hardware-based system in this thesis is capable of effectively minimizing reflected sound waves from an underwater surface. This outstanding feature could be used in future applications to conceal underwater vehicles from active sonar. Overall, the findings in this thesis provide compelling arguments for hardware-based adaptive filters for acoustic applications
In this thesis, acoustic applications of adaptive filters implemented on Field Programmable Gate Arrays will be presented and evaluated. The developed systems in this work include well-known setups, such as Kundt’s Tube or an active headrest, for validation purposes. The more advanced solutions incorporate a helicopter headset with additional speech enhancement capabilities and active impedance control of an underwater surface. Additionally, an overview of the theoretical background is given, and state-of-the-art related work is discussed.
The results of the measurements show that hardware-based adaptive filters not only match their software counterparts, but they can enhance noise reduction performance. In the case of the helicopter headset, the additional computational headroom of the chosen platform is used to include Wiener filter-based speech enhancement. This system produces impressive results in the evaluated speech quality, in addition to the active noise reduction of the headset. Active control of underwater sound bears additional challenges when it comes to latency requirements due to the high speed of sound. The proposed hardware-based system in this thesis is capable of effectively minimizing reflected sound waves from an underwater surface. This outstanding feature could be used in future applications to conceal underwater vehicles from active sonar. Overall, the findings in this thesis provide compelling arguments for hardware-based adaptive filters for acoustic applications
Version
Published version
Access right on openHSU
Open access