Design of Data Driven Automated Driving Control Algorithm for Enhanced Human Acceptance

This article presents a data-driven automated driving control algorithm designed to enhance human acceptance in autonomous vehicles. To achieve this, we utilize an LSTM autoencoder to extract latent driving features from collected data, which are then clustered into lateral and longitudinal behavior...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on consumer electronics Jg. 71; H. 3; S. 7848 - 7863
Hauptverfasser: Shin, Donghoon, Han, YongHa, Park, Kang-Moon
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York IEEE 01.08.2025
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Schlagworte:
Acceptance
Algorithms
autoencoder
Autoencoders
Automation
Autonomous driving
Autonomous vehicles
Control algorithms
Control theory
Data collection
Data mining
data visualization
deep learning
Discomfort
Driver behavior
driving behavior
Feature extraction
human factor
Long short term memory
Passenger comfort
Quality assessment
ride comfort
Riding quality
risk assessment
Roads
Safety
Trajectory
Vehicles
ISSN:0098-3063, 1558-4127
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This article presents a data-driven automated driving control algorithm designed to enhance human acceptance in autonomous vehicles. To achieve this, we utilize an LSTM autoencoder to extract latent driving features from collected data, which are then clustered into lateral and longitudinal behaviors. These clustered behaviors serve as the foundation for generating steering and acceleration profiles, enabling the automated driving system to replicate individualized human driving styles. By incorporating three distinct driving behaviors, the proposed approach effectively mitigates potential sources of discomfort, such as excessive jerk and abrupt accelerations, ensuring a smoother ride experience. Additionally, we introduce a comprehensive ride quality evaluation metric that considers both trajectory similarity (trajectory score) and passenger comfort (comfort energy expression). The effectiveness of the proposed algorithm is validated through real-world vehicle tests, focusing on driving scenarios known to cause ride discomfort. Experimental results demonstrate that the automated driving control framework successfully enhances ride quality while adapting to individual passenger driving preferences, thus improving overall human acceptance of autonomous vehicles.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0098-3063
1558-4127
DOI:10.1109/TCE.2025.3597297