Hybrid CNN Architectures for Speech Emotion Recognition in Intelligent Cockpit Environments
ITM Web of Conferences · 2026 · 93 · 03003
DOI: 10.1051/itmconf/20269303003 ↗Kurz gefasst
Fachartikel in der Zeitschrift ITM Web of Conferences, online erschienen am 13.11.2026. Verfasst von Zhisheng Zhang. Die Arbeit nennt 11 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
With the rapid development of intelligent connected vehicles, speech emotion recognition has become a key technology to enhance human‑machine interaction experience and ensure driving safety, yet it still faces severe challenges such as strong in‑cabin non‑stationary noise, limited on‑board computing resources, and scarce labeled emotional datasets. This paper focuses on analyzing the intrinsic mechanisms, structural advantages and applicable scenarios of representative models including one‑dimensional CNN, CNN‑LSTM, Transformer‑CNN and MFA‑CNN, and explores the lightweight implementation value of 2D‑3D convolution fusion and neural architecture search technology in real‑time vehicle‑side deployment. On this basis, this paper further discusses the application logic and engineering value of multimodal feature fusion, NPU‑CPU heterogeneous inference and human‑machine collaboration mechanisms in complex cockpit scenes, so as to improve recognition robustness and response efficiency. Finally, this article summarizes the current research bottlenecks in noise robustness, computing power‑precision‑delay balance, data scarcity and cross‑modal alignment, and provides a forward‑looking outlook on the engineering implementation path of lightweight model design, model compression, edge‑cloud collaboration and privacy‑preserving computing, aiming to provide theoretical reference and technical guidance for the industrial application of speech emotion recognition in next‑generation intelligent cockpits.
Zitieren
APA
Zhang, Z. (2026). Hybrid CNN Architectures for Speech Emotion Recognition in Intelligent Cockpit Environments. ITM Web of Conferences, 93, 03003. https://doi.org/10.1051/itmconf/20269303003
Vancouver
Zhang Z. Hybrid CNN Architectures for Speech Emotion Recognition in Intelligent Cockpit Environments. ITM Web of Conferences. 2026;93:03003. doi:10.1051/itmconf/20269303003
BibTeX
@article{Zhang2026,
author = {Zhang, Zhisheng},
title = {Hybrid CNN Architectures for Speech Emotion Recognition in Intelligent Cockpit Environments},
journal = {ITM Web of Conferences},
year = {2026},
volume = {93},
pages = {03003},
doi = {10.1051/itmconf/20269303003}
}
EUVIMED