Understanding and Implementing Smart Home Voice Commands and Intent Recognition Based on Deep Learning
Year: 2025
Type: article
Abstract: With the flourishing development of the Internet of Things and artificial intelligence, smart home devices have become increasingly popular. However, accurately recognizing and understanding users' voice commands and intentions in smart home scenarios remains a challenging task. This paper aims to construct a model and system that can precisely interpret smart home users' voice commands and intentions through deep learning technology. It adopts methods such as data cleaning, data augmentation, and model construction (including a CNN-Transformer-based speech recognition model and a fine-tuned BERT-based natural language processing model). The study reveals that the proposed models outperform traditional models in terms of accuracy, recall, and F1-score. This research is of great significance for enhancing user experience, promoting the development of the smart home industry, and facilitating technological innovation in related fields.
Author: Meng Tian
Institution: Harbin Institute of Technology
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4410129506