Visionspace is a company that develops and integrates enterprise-level solutions for satellite missions. They are known for being a reliable and experienced partner in the monitoring of satellites and ground stations. Their systems leverage artificial intelligence and automation to manage complex control systems, leading to successful missions.

Visionspace recognized an opportunity to optimize internal communication processes. They aimed to improve the accessibility and usability of voice loops, which are discrete audio channels used for information exchange in spacecraft operations. While voice loops provided valuable insights into the procedures of spacecraft operations, the current paradigm of storing them as raw audio files made it challenging to extract the information contained within. Visionspace saw the potential to leverage artificial intelligence and automation to improve their operations and maximize the benefits of voice loop data.

The challenge:

*Extracting information from raw audio files in voice loops was time-consuming and labor-intensive.

*The technical language and mix of accents in spacecraft operations required a customized solution for accurate transcriptions.

*The lack of searchable text hindered efficient analysis and review during intensive operational periods.

Solution:

*Developed an Automatic Speech Recognition (ASR) system to index transcriptions of audio files.

*ASR prototype consisted of:

A client Raspberry Pi interfacing with ESOC's voice loop system for recording and

transmitting audio.

A backend server utilizing a state-of-the-art speech-to-text engine for audio

transcription and storage.

A frontend dashboard UI enabling operators to view, search, and edit

transcriptions.

*Customized the ASR system by training it with specific language and domain data to improve accuracy.

Results:

*Transcriptions became easily understandable to human readers, enhancing accessibility and usability of voice loop data.

*Accessible and searchable transcriptions facilitated efficient error analysis and quality assurance processes.

*Operators could easily search and review specific audio communications using keywords and date-time references, reducing manual effort and saving time.

*The system provided valuable insights into spacecraft procedures, leading to improved operational decision-making and mission success.

*The scalability of the ASR system allowed for seamless integration with multiple missions, enhancing overall operational efficiency.

More: https://1d.works/case-studies/vision-space