Large Language Model-based Intent Parsing Technology for Automotive Voice Commands

Li Yinan

School of Computer Information and Engineering, NJTECH University

Abstract：

This paper studies an intent parsing technology for automotive voice commands based on a large language model. This technology effectively improves the accuracy and real-time response of natural language understanding through Transformer self-attention mechanism, cueing engineering, model fine-tuning, and multimodal fusion. It overcomes the limitations of traditional methods in handling ambiguity, context dependence, and noisy environments, significantly reduces the cognitive load on drivers, and enhances the naturalness of interaction. Related research shows that it exhibits significant advantages in intent recognition accuracy and system robustness, providing important support for the upgrade of human-machine interfaces in intelligent vehicles.

Key Words：

large language model; automotive voice interaction; intent parsing; prompting engineering; multimodal fusion