Skip to main content

Call for Papers: Signal Processing and Machine Learning for Speech and Audio in Acoustic Sensor Networks

Besides the omnipresent mobile phones, we are surrounded by many recording devices such as laptop computers, tablets, smart watches, camcorders, and others. While many speech and audio applications have been traditionally implemented on compact devices featuring one or more microphones, wireless acoustic sensor networks (WASN) now offer a new paradigm for acoustic sensing and processing, bearing the promise to overcome the limitations of individual devices. The integration of WASNs in new speech and audio applications triggers a host of challenging research questions, solutions to which are at the core of this special issue.

As WASNs enable a distributed configuration of recording devices, efficient and robust solutions for handling asynchronous channels, unknown microphone positions, and distributed computing platforms are urgently sought. Appropriate methods for tackling these challenges range from traditional acoustic signal processing and statistical detection and estimation to data-driven non-linear networks, where for the sake of computational efficiency a blend of these methods in the form of model-based deep learning approaches appears especially attractive. The development of these methods will in turn require novel data sets and simulation methods where also interactions with wireless communications may play an important role. As a consequence, significant improvements in the performance of various audio tasks such as speech enhancement, speech recognition, diarization, scene analysis, and anomalous acoustic event detection may be expected. In fact, acquiring acoustic information with sensors distributed over large spaces and exploiting the recent advances in machine learning has already shown to significantly contribute to emerging technologies for smart homes, autonomous driving, and environmental monitoring. To this end, we invite authors to submit papers that demonstrate challenging research questions, methods and solutions for dealing with ad-hoc distributed microphone signals and applications thereof.

The topics of interest for the special issue include, but are not limited to:

  • Sensor selection, self-localization, clustering and synchronization
  • Signal enhancement, extraction, and separation
  • Acoustic source localization and classification
  • Acoustic scene analysis
  • Synergies with multimodal systems
  • Privacy and security in WASNs
  • Computationally efficient and robust federated approaches
  • Model-based deep-learning approaches for WASNs

Important Dates
Submission deadline: March 15, 2023

Lead Guest Editor
Walter Kellermann, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany

Guest Editors
Nobutaka Ono, Tokyo Metropolitan University, Japan
Rainer Martin, Ruhr-Universität Bochum, Germany

Important: Authors should select "Signal Processing and Machine Learning for Speech and Audio in Acoustic Sensor Networks" when they reach the Thematic Series section in the submission system.

Submission guidelines