CADRE: A Cascade Deep Reinforcement Learning Framework for Vision-based Autonomous Urban Driving

2022/02/17 by Yinuo Zhao, Zhao, Yinuo, Kun Wu +11 · 5 citations
Computer Science · Engineering · Psychology · #Advanced Neural Network Applications #Artificial Intelligence (cs.AI) #Artificial intelligence #Autonomous Vehicle Technology and Safety #Benchmark (surveying) #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Deep learning #FOS: Computer and information sciences #Human–computer interaction #Machine learning #Margin (machine learning) #Mobile robot #Obstacle avoidance #Perception #Psychology #Reinforcement learning #Robot #Video Surveillance and Tracking Methods

paper · pdf · doi:10.48550/arxiv.2202.08557

published in arXiv (Cornell University) (Cornell University)

openalex publication_date 2022/02/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/06

Abstract

Vision-based autonomous urban driving in dense traffic is quite challenging due to the complicated urban environment and the dynamics of the driving behaviors. Widely-applied methods either heavily rely on hand-crafted rules or learn from limited human experience, which makes them hard to generalize to rare but critical scenarios. In this paper, we present a novel CAscade Deep REinforcement learning framework, CADRE, to achieve model-free vision-based autonomous urban driving. In CADRE, to derive representative latent features from raw observations, we first offline train a Co-attention Perception Module (CoPM) that leverages the co-attention mechanism to learn the inter-relationships between the visual and control information from a pre-collected driving dataset. Cascaded by the frozen CoPM, we then present an efficient distributed proximal policy optimization framework to online learn the driving policy under the guidance of particularly designed reward functions. We perform a comprehensive empirical study with the CARLA NoCrash benchmark as well as specific obstacle avoidance scenarios in autonomous urban driving tasks. The experimental results well justify the effectiveness of CADRE and its superiority over the state-of-the-art by a wide margin.

Cited by

• Advancing Autonomous Driving System Testing: Demands, Challenges, and Future Directions
• Collaborative Scheduling of Time-dependent UAVs,Vehicles and Workers for Crowdsensing in Disaster Response
• Action Robust Reinforcement Learning via Optimal Adversary Aware Policy Optimization
• Autonomous Collaborative Scheduling of Time-dependent UAVs, Workers and Vehicles for Crowdsensing in Disaster Response
• Offline Reinforcement Learning using Human-Aligned Reward Labeling for Autonomous Emergency Braking in Occluded Pedestrian Crossing

Related

• Interpretable End-to-end Urban Autonomous Driving with Latent Deep Reinforcement Learning
• Vision-Based Autonomous Driving: A Model Learning Approach
• Model-free Deep Reinforcement Learning for Urban Autonomous Driving
• Efficient Learning of Urban Driving Policies Using Bird's-Eye-View State Representations
• WAD: A Deep Reinforcement Learning Agent for Urban Autonomous Driving