Date Approved

9-23-2022

Embargo Period

9-27-2022

Document Type

Thesis

Degree Name

M.S. Electrical and Computer Engineering

Department

Electrical and Computer Engineering

College

Henry M. Rowan College of Engineering

Advisor

Ravi P. Ramachandran, Ph.D.

Committee Member 1

Parth Bhavsar, Ph.D.

Committee Member 2

Yusuf Mehta, Ph.D., P.E.

Committee Member 3

John Schmalzel, Ph.D., P.E.

Subject(s)

Automatic speech recognition; Air traffic control; Machine learning

Disciplines

Artificial Intelligence and Robotics | Electrical and Computer Engineering | Multi-Vehicle Systems and Air Traffic Control

Abstract

Advances in Artificial Intelligence and Machine learning have enabled a variety of new technologies. One such technology is Automatic Speech Recognition (ASR), where a machine is given audio and transcribes the words that were spoken. ASR can be applied in a variety of domains to improve general usability and safety. One such domain is Air Traffic Control (ATC). ASR in ATC promises to improve safety in a mission critical environment. ASR models have historically required a large amount of clean training data. ATC environments are noisy and acquiring labeled data is a difficult, expertise dependent task. This thesis attempts to solve these problems by presenting a machine learning framework which uses word-by-word audio samples to transcribe ATC speech. Instead of transcribing an entire speech sample, this framework transcribes every word individually. Then, overall transcription is pieced together based on the word sequence. Each stage of the framework is trained and tested independently of one another, and the overall performance is gauged. The overall framework was gauged to be a feasible approach to ASR in ATC.

Download

Share

COinS