ASTRA: A Scalable Next-Generation ATCO Training Simulator with Autonomous Simpilots

2026-06-16

AI Summary

This study addresses the limitations of current air traffic control officer (ATCO) training, which relies on human role-players as simulated pilots (simpilots), and the poor performance of generic automatic speech recognition (ASR) systems on Singapore-accented aviation speech—exhibiting a word error rate (WER) as high as 107.80%. To overcome these challenges, the authors propose ASTRA, an end-to-end automated training simulator that integrates an ASR model fine-tuned for local accents alongside a tailored text-to-speech module. ASTRA further incorporates semantic understanding and response generation components powered by DSPy and Unsloth, establishing the first fully automated simpilot system designed specifically for Singapore’s aviation environment. A multidimensional AI-based evaluation framework is introduced to quantitatively assess response accuracy, conciseness, and completeness. Experimental results demonstrate a substantial reduction in WER to 23.45%, with evaluation scores reaching 91.7%, 88.2%, and 86.9% respectively, significantly enhancing training standardization and scalability.

Abstract

Air Traffic Control Operators (ATCOs) are vital in ensuring the safe, orderly, and efficient flow of air traffic, yet training capacity is constrained by reliance on specialized human trainers known as simpilots, who must role-play both pilots and ATCOs in a simulated airspace. Existing automated solutions rely on Western-centric speech models that perform poorly in Singaporean operational contexts, with off-the-shelf systems exhibiting Word Error Rates (WER) of up to 107.80% on Singaporean-accented aviation speech. We introduce ASTRA, an end-to-end training simulator that automates these simpilot roles through a pipeline that transcribes ATCO speech, interprets instructions, and generates appropriate pilot and ATCO responses using locally adapted voice models. Our fine-tuned Automatic Speech Recognition (ASR) pipeline reduces WER to 23.45%, substantially outperforming existing approaches in this domain. Beyond traffic simulation, ASTRA incorporates an AI-assisted performance evaluation framework that assesses trainee radiotelephony communications across accuracy, brevity, and completeness, achieving post-optimization scores of 91.7%, 88.2%, and 86.9%, respectively. Built on open-source foundations such as DSPy and Unsloth, this approach enables scalable, standardized ATCO assessment while reducing instructor workload.

Problem

• Air Traffic Control
• ATCO training
• speech recognition
• accent robustness
• simulation scalability

Innovation

• ASTRA
• Automatic Speech Recognition
• Simpilot Automation
• AI-assisted Evaluation
• Accent-adapted Voice Models