Saliency-Guided Domain Adaptation for Left-Hand Driving in Autonomous Steering

2025-11-02

AI Summary

This study addresses the poor generalization of autonomous driving models (e.g., PilotNet, ResNet) trained on right-hand-drive data when deployed in left-hand-drive environments (e.g., Australia). We propose a domain adaptation method comprising image-flip pretraining followed by fine-tuning on real left-hand-drive data. Crucially, we introduce saliency map analysis to quantitatively validate the effective shift of model attention from right-side driving cues to critical left-side regions. Experiments demonstrate that our approach significantly reduces steering prediction error—by 23.6% relative to the baseline—while enhancing stability and robustness in left-hand-drive scenarios. Unlike conventional fine-tuning or adversarial alignment methods, ours requires no additional annotations or complex architectural components, achieving both computational efficiency and interpretability. This work establishes a lightweight, principled paradigm for cross-driving-rule domain adaptation in autonomous driving systems.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Adversarial Attacks & Robustness
• Application Domains: Mobility, Driving & Flight

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Domain adaptation is required for automated driving models to generalize well across diverse road conditions. This paper explores a training method for domain adaptation to adapt PilotNet, an end-to-end deep learning-based model, for left-hand driving conditions using real-world Australian highway data. Four training methods were evaluated: (1) a baseline model trained on U.S. right-hand driving data, (2) a model trained on flipped U.S. data, (3) a model pretrained on U.S. data and then fine-tuned on Australian highways, and (4) a model pretrained on flipped U.S. data and then finetuned on Australian highways. This setup examines whether incorporating flipped data enhances the model adaptation by providing an initial left-hand driving alignment. The paper compares model performance regarding steering prediction accuracy and attention, using saliency-based analysis to measure attention shifts across significant road regions. Results show that pretraining on flipped data alone worsens prediction stability due to misaligned feature representations, but significantly improves adaptation when followed by fine-tuning, leading to lower prediction error and stronger focus on left-side cues. To validate this approach across different architectures, the same experiments were done on ResNet, which confirmed similar adaptation trends. These findings emphasize the importance of preprocessing techniques, such as flipped-data pretraining, followed by fine-tuning to improve model adaptation with minimal retraining requirements.

Problem

• Adapting autonomous steering models to left-hand driving conditions
• Evaluating domain adaptation methods using flipped data preprocessing
• Analyzing steering prediction accuracy and attention shifts across domains

Innovation

• Saliency-guided domain adaptation for autonomous steering
• Flipped-data pretraining followed by fine-tuning
• Attention shift analysis using saliency-based methods