mlsanityData to 7 Oct 2026

Improved Reinforcement Learning through Imitation Learning Pretraining Towards Image-based Autonomous Driving

arXiv 1907.06838 ↗First posted 16 Jul 2019cs.LG cs.AI cs.CVcited by 5

On the map: Offline RL and imitation learning from data → Imitation and offline RL for driving

Abstract (arXiv)

We present a training pipeline for the autonomous driving task given the current camera image and vehicle speed as the input to produce the throttle, brake, and steering control output. The simulator Airsim's convenient weather and lighting API provides a sufficient diversity during training which can be very helpful to increase the trained policy's robustness. In order to not limit the possible policy's performance, we use a continuous and deterministic control policy setting. We utilize ResNet-34 as our actor and critic networks with some slight changes in the fully connected layers. Considering human's mastery of this task and the high-complexity nature of this task, we first use imitation learning to mimic the given human policy and leverage the trained policy and its weights to the reinforcement learning phase for which we use DDPG. This combination shows a considerable performance boost comparing to both pure imitation learning and pure DDPG for the autonomous driving task.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Placed in Imitation and offline RL for driving, one of 13 directions in Offline RL and imitation learning from data. Builds on 13 of its 15 references that are in the map. First posted on arXiv 16 Jul 2019.

Builds on

13 of 15 references.

References · 13

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  3. 2017Squeeze-and-Excitation Networkscited by 4,763
  4. 2015Continuous control with deep reinforcement learningcited by 3,279
  5. 2018CBAM: Convolutional Block Attention Modulecited by 2,463
  6. 2013Rich feature hierarchies for accurate object detection and semantic segmentationcited by 2,160
  7. 2017AirSim: High-Fidelity Visual and Physical Simulation for Autonomous Vehiclescited by 517
  8. 2016CAD2RL: Real Single-Image Flight without a Single Real Imagecited by 227
  9. 2018BAM: Bottleneck Attention Modulecited by 171
  10. 2017Towards Monocular Vision based Obstacle Avoidance through Deep Reinforcement Learningcited by 20
  11. 2019Learning Accurate, Comfortable and Human-like Drivingcited by 8
  12. 2018Learning Transferable UAV for Forest Visual Perceptioncited by 3
1 more
  1. 2019Drone Path-Following in GPS-Denied Environments using Convolutional Networkscited by 2

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 0; 2023: 0; 2024: 2; 2025: 2.

5 papers.

  1. 2024DriveCoT: Integrating Chain-of-Thought Reasoning with End-to-End Drivingcited by 67
  2. 2024QADQN: Quantum Attention Deep Q-Network for Financial Market Predictioncited by 8
  3. 2025Emoti-Attack: Zero-Perturbation Adversarial Attacks on NLP Systems via Emoji Sequencescited by 5
  4. 2021Integrating Imitation Learning with Human Driving Data into Reinforcement Learning to Improve Training Efficiency for Autonomous Drivingcited by 0
  5. 2025Learning-based Autonomous Oversteer Control and Collision Avoidancecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2021Improved Deep Reinforcement Learning with Expert Demonstrations for Urban Autonomous Driving
  2. 2021Generative Adversarial Imitation Learning for End-to-End Autonomous Driving on Urban Environments
  3. 2022Imitation Is Not Enough: Robustifying Imitation with Reinforcement Learning for Challenging Driving Scenarios
  4. 2021Autonomous Racing using a Hybrid Imitation-Reinforcement Learning Architecture
  5. 2026Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving
  6. 2022Imitation Learning for Generalizable Self-driving Policy with Sim-to-real Transfer
  7. 2021Vision-Based Autonomous Car Racing Using Deep Imitative Reinforcement Learning
  8. 2020Learning to Drive (L2D) as a Low-Cost Benchmark for Real-World Reinforcement Learning