Energy-Efficient Autonomous Driving with Adaptive Perception and Robust Decision

2025-10-29

AI Summary

To address the high energy consumption of perception computation in autonomous driving—which critically limits the range of electric vehicles—this paper proposes EneAD, a framework that jointly optimizes energy efficiency and driving performance. Methodologically, EneAD integrates three key innovations: (1) a transferable Bayesian optimization mechanism that dynamically adjusts frame rates and computational intensity across multiple perception models; (2) a lightweight scene classifier enabling perception difficulty–aware adaptation; and (3) a regularized reinforcement learning–based decision module to enhance robustness against perception perturbations. Extensive experiments demonstrate that, while maintaining safety-critical driving performance, EneAD reduces perception energy consumption by 1.9–3.5× and extends vehicle range by 3.9%–8.5%, significantly outperforming state-of-the-art model compression and adaptive perception approaches.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Autonomous driving is an emerging technology that is expected to bring significant social, economic, and environmental benefits. However, these benefits come with rising energy consumption by computation engines, limiting the driving range of vehicles, especially electric ones. Perception computing is typically the most power-intensive component, as it relies on largescale deep learning models to extract environmental features. Recently, numerous studies have employed model compression techniques, such as sparsification, quantization, and distillation, to reduce computational consumption. However, these methods often result in either a substantial model size or a significant drop in perception accuracy compared to high-computation models. To address these challenges, we propose an energy-efficient autonomous driving framework, called EneAD. In the adaptive perception module, a perception optimization strategy is designed from the perspective of data management and tuning. Firstly, we manage multiple perception models with different computational consumption and adjust the execution framerate dynamically. Then, we define them as knobs and design a transferable tuning method based on Bayesian optimization to identify promising knob values that achieve low computation while maintaining desired accuracy. To adaptively switch the knob values in various traffic scenarios, a lightweight classification model is proposed to distinguish the perception difficulty in different scenarios. In the robust decision module, we propose a decision model based on reinforcement learning and design a regularization term to enhance driving stability in the face of perturbed perception results. Extensive experiments evidence the superiority of our framework in both energy consumption and driving performance. EneAD can reduce perception consumption by 1.9x to 3.5x and thus improve driving range by 3.9% to 8.5%

Problem

• Reducing high energy consumption in autonomous driving perception systems
• Maintaining perception accuracy while minimizing computational requirements
• Ensuring robust decision-making under varying traffic conditions

Innovation

• Adaptive perception with dynamic model and framerate tuning
• Lightweight classification for scenario-based perception adjustment
• Reinforcement learning decision model with stability regularization