Comparison of AI Speech-to-Text Systems and Their Application in Artillery Command and Fire Control Systems

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60162694%3AG42__%2F26%3A00564679" target="_blank" >RIV/60162694:G42__/26:00564679 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.springer.com/series/7899" target="_blank" >https://www.springer.com/series/7899</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-032-04339-9_6" target="_blank" >10.1007/978-3-032-04339-9_6</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Comparison of AI Speech-to-Text Systems and Their Application in Artillery Command and Fire Control Systems

Popis výsledku v původním jazyce

This paper presents a comparative analysis of three leading AI speech-to-text (STT) systems: Descript.com, Google Vertex AI Studio (Chirp), and OpenAI Whisper. The objective of the study is to evaluate the accuracy, functionality, and potential applications of these technologies, with a particular focus on their integration into artillery command and fire control systems. The analysis outlines the evolution of speech recognition technologies, from traditional methods based on Hidden Markov Models (HMMs) to modern deep neural networks, including Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs), and Transformer-based architectures. Practical testing was conducted on a dataset of English and Czech recordings with varying audio quality. The results indicate that Google Chirp achieves the highest accuracy in English transcriptions, while OpenAI Whisper demonstrates superior performance for the Czech language. Additionally, the paper explores the optimization of STT systems for combat environments, including the use of Ant Colony Optimization (ACO) algorithms to minimize errors and enhance the relevance of transcriptions. The study also highlights security risks associated with deploying cloud-based STT services in military applications and emphasizes the advantages of on-premise solutions to ensure data protection. Finally, the paper discusses strategies for modernizing defense capabilities through AI technologies. It advocates for increased investment in automated command and control systems, fire control, and situational awareness, emphasizing their crucial role in improving response times and the accuracy of artillery fire support.

Název v anglickém jazyce

Comparison of AI Speech-to-Text Systems and Their Application in Artillery Command and Fire Control Systems

Popis výsledku anglicky

This paper presents a comparative analysis of three leading AI speech-to-text (STT) systems: Descript.com, Google Vertex AI Studio (Chirp), and OpenAI Whisper. The objective of the study is to evaluate the accuracy, functionality, and potential applications of these technologies, with a particular focus on their integration into artillery command and fire control systems. The analysis outlines the evolution of speech recognition technologies, from traditional methods based on Hidden Markov Models (HMMs) to modern deep neural networks, including Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs), and Transformer-based architectures. Practical testing was conducted on a dataset of English and Czech recordings with varying audio quality. The results indicate that Google Chirp achieves the highest accuracy in English transcriptions, while OpenAI Whisper demonstrates superior performance for the Czech language. Additionally, the paper explores the optimization of STT systems for combat environments, including the use of Ant Colony Optimization (ACO) algorithms to minimize errors and enhance the relevance of transcriptions. The study also highlights security risks associated with deploying cloud-based STT services in military applications and emphasizes the advantages of on-premise solutions to ensure data protection. Finally, the paper discusses strategies for modernizing defense capabilities through AI technologies. It advocates for increased investment in automated command and control systems, fire control, and situational awareness, emphasizing their crucial role in improving response times and the accuracy of artillery fire support.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Communications in Computer and Information Science

ISBN

978-3-032-04338-2

ISSN

1865-0929

e-ISSN

1865-0937

Počet stran výsledku

14

Strana od-do

82-95

Název nakladatele

SPRINGER INTERNATIONAL PUBLISHING AG

Místo vydání

Bilbao

Místo konání akce

Bilbao, SPAIN

Datum konání akce

12. 6. 2025

Typ akce podle státní příslušnosti

CST - Celostátní akce

Kód UT WoS článku

001696487100006