Operation Language Acquisition for Scenarios: A BERT-Driven Augmented Reality System for Defense English Competence

Abstract

As operation language acquisition tasks have increasingly higher requirements for real operation, scenario-driven and intelligent use, current operational English acquisition methods have such problems like delayed response and poor adaptability in terms of context perception, feedback timeliness and immersive interactivity. To this end, this paper introduces the integration of artificial intelligence and augmented reality (AR) technology, and realizes semantic understanding, instant dialogue generation and interactive feedback in tactical tasks by building a natural language processing framework driven on the BERT model and coordinating AR visualization and speech recognition systems. The specific method includes fine-tuning the BERT model to adapt to the wartime language, combining AR head-mounted displays and voice input modules to present command contexts and semantic responses in real time, and improving learners' language processing capabilities in realistic task scenarios. The results show that the experimental group outperforms the control group in both recognition accuracy and feedback efficiency, verifying the technical advantages of the improved teaching system in voice interaction. In terms of speech recognition accuracy, the experimental group scored 98.5%, 97%, 98.2% and 99% in the four tasks, respectively, with an average accuracy of 98.18%, verifying the effectiveness and advancement of the fusion system in detail control and scene adaptability.