Publication Details

Grammar Based Speaker Role Identification for Air Traffic Control Speech Recognition

PRASAD Amrutha, ZULUAGA-GOMEZ Juan, MOTLÍČEK Petr, SARFJOO Seyyed Saeed, NIGMATULINA Iuliia, OHNEISER Oliver and HELMKE Hartmut. Grammar Based Speaker Role Identification for Air Traffic Control Speech Recognition. In: Proceedings of the 12th SESAR Innovation Days. Budapest, 2022, pp. 1-9. Available from: https://arxiv.org/abs/2108.12175
Czech title
Identifikace role mluvčího pro rozpoznávání řeči při řízení letového provozu na základě gramatiky
Type
conference paper
Language
english
Authors
Prasad Amrutha (DCGM FIT BUT)
Zuluaga-Gomez Juan (IDIAP)
Motlíček Petr, doc. Ing., Ph.D. (DCGM FIT BUT)
Sarfjoo Seyyed Saeed (IDIAP)
Nigmatulina Iuliia (IDIAP)
Ohneiser Oliver (DLR)
Helmke Hartmut (DLR)
URL
Keywords

assistant based speech recognition, air traffic management, multitask acoustic modeling, speaker role classification, Kaldi

Abstract

Automatic Speech Recognition (ASR) for air traffic control is generally trained by pooling Air Traffic Controller (ATCO) and pilot data into one set. This is motivated by the fact that pilot's voice communications are more scarce than ATCOs. Due to this data imbalance and other reasons (e.g., varying acoustic conditions), the speech from ATCOs is usually recognized more accurately than from pilots. Automatically identifying the speaker roles is a challenging task, especially in the case of the noisy voice recordings collected using Very High Frequency (VHF) receivers or due to the unavailability of the push-to-talk (PTT) signal, i.e., both audio channels are mixed. In this work, we propose to (1) automatically segment the ATCO and pilot data based on an intuitive approach exploiting ASR transcripts and (2) subsequently consider an automatic recognition of ATCOs' and pilots' voice as two separate tasks. Our work is performed on VHF audio data with high noise levels, i.e., signal-to-noise (SNR) ratios below 15 dB, as this data is recognized to be helpful for various speech-based machinelearning tasks. Specifically, for the speaker role identification task, the module is represented by a simple yet efficient knowledge-based system exploiting a grammar defined by the International Civil Aviation Organization (ICAO). The system accepts text as the input, either manually verified annotations or automatically generated transcripts. The developed approach provides an average accuracy in speaker role identification of about 83%. Finally, we show that training an acoustic model for ASR tasks separately (i.e., separate models for ATCOs and pilots) or using a multitask approach is well suited for the noisy data and outperforms the traditional ASR system where all data is pooled together.

Published
2022
Pages
1-9
Proceedings
Proceedings of the 12th SESAR Innovation Days
Conference
12th SESAR Innovation Days, Budapešť, HU
Place
Budapest, HU
BibTeX 
@INPROCEEDINGS{FITPUB13049,
   author = "Amrutha Prasad and Juan Zuluaga-Gomez and Petr Motl\'{i}\v{c}ek and Saeed Seyyed Sarfjoo and Iuliia Nigmatulina and Oliver Ohneiser and Hartmut Helmke",
   title = "Grammar Based Speaker Role Identification for Air Traffic Control Speech Recognition",
   pages = "1--9",
   booktitle = "Proceedings of the 12th SESAR Innovation Days",
   year = 2022,
   location = "Budapest, HU",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/13049"
}
Back to top