CARIL: Confidence-Aware Regression in Imitation Learning for Autonomous Driving

2025-03-02

AI Summary

To address the challenge of quantifying and correcting action reliability in end-to-end vision-based autonomous driving, this paper proposes a regression-classification dual-head network architecture. The regression head outputs continuous control commands, while the classification head—novelly employed for confidence modeling in regression tasks—enables real-time estimation of prediction trustworthiness. Confidence scores dynamically trigger closed-loop action correction. This approach overcomes two fundamental limitations of prior methods: the absence of uncertainty quantification in conventional regression models and the accuracy degradation inherent to classification-based approaches due to output discretization. Evaluated in the CARLA simulator, the method reduces lane departure rate by 50% and achieves significantly higher trajectory accuracy than pure regression baselines, thereby jointly optimizing both control precision and operational reliability.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Learning & Optimization for CV
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

End-to-end vision-based imitation learning has demonstrated promising results in autonomous driving by learning control commands directly from expert demonstrations. However, traditional approaches rely on either regressionbased models, which provide precise control but lack confidence estimation, or classification-based models, which offer confidence scores but suffer from reduced precision due to discretization. This limitation makes it challenging to quantify the reliability of predicted actions and apply corrections when necessary. In this work, we introduce a dual-head neural network architecture that integrates both regression and classification heads to improve decision reliability in imitation learning. The regression head predicts continuous driving actions, while the classification head estimates confidence, enabling a correction mechanism that adjusts actions in low-confidence scenarios, enhancing driving stability. We evaluate our approach in a closed-loop setting within the CARLA simulator, demonstrating its ability to detect uncertain actions, estimate confidence, and apply real-time corrections. Experimental results show that our method reduces lane deviation and improves trajectory accuracy by up to 50%, outperforming conventional regression-only models. These findings highlight the potential of classification-guided confidence estimation in enhancing the robustness of vision-based imitation learning for autonomous driving. The source code is available at https://github.com/ElaheDlv/Confidence_Aware_IL.

Problem

• Combines regression and classification for reliable autonomous driving decisions.
• Enables real-time action correction in low-confidence driving scenarios.
• Reduces lane deviation and improves trajectory accuracy by 50%.

Innovation

• Dual-head neural network for driving actions
• Confidence estimation with classification head
• Real-time correction in low-confidence scenarios