mlsanityData to 14 Oct 2026

Dream to Fly: Model-Based Reinforcement Learning for Vision-Based Drone Flight

arXiv 2501.14377 ↗First posted 24 Jan 2025cs.ROcited by 15

On the map: Deep reinforcement learning methods → RL for robot navigation and UAV control

Abstract (arXiv)

Autonomous drone racing has risen as a challenging robotic benchmark for testing the limits of learning, perception, planning, and control. Expert human pilots are able to fly a drone through a race track by mapping pixels from a single camera directly to control commands. Recent works in autonomous drone racing attempting direct pixel-to-commands control policies have relied on either intermediate representations that simplify the observation space or performed extensive bootstrapping using Imitation Learning (IL). This paper leverages DreamerV3 to train visuomotor policies capable of agile flight through a racetrack using only pixels as observations. In contrast to model-free methods like PPO or SAC, which are sample-inefficient and struggle in this setting, our approach acquires drone racing skills from pixels. Notably, a perception-aware behaviour of actively steering the camera toward texture-rich gate regions emerges without the need of handcrafted reward terms for the viewing direction. Our experiments show in both, simulation and real-world flight using a hardware-in-the-loop setup with rendered image observations, how the proposed approach can be deployed on real quadrotors at speeds of up to 9 m/s. These results advance the state of pixel-based autonomous flight and demonstrate that MBRL offers a promising path for real-world robotics research.

Key facts. Cited by 15 papers in the map (data through 14 Oct 2026). Placed in RL for robot navigation and UAV control, one of 15 directions in Deep reinforcement learning methods. Builds on 45 of its 55 references that are in the map. First posted on arXiv 24 Jan 2025.

Builds on

45 of 55 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024Multi-Task Reinforcement Learning for Quadrotors17 Dec
  2. 2024Demonstrating Agile Flight from Pixels without State Estimation18 Jun
  3. 2024Bootstrapping Reinforcement Learning with Imitation for Vision-Based Agile Flight18 Mar

Other references · 42

  1. 2017Proximal Policy Optimization Algorithmscited by 15,215
  2. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
  3. 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 1,653
  4. 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,498
  5. 2018Learning Latent Dynamics for Planning from Pixelscited by 1,264
  6. 2023Mastering Diverse Domains through World Modelscited by 1,253
  7. 2019Habitat: A Platform for Embodied AI Researchcited by 1,191
  8. 2018DeepMind Control Suitecited by 956
  9. 2015End-to-End Training of Deep Visuomotor Policiescited by 875
  10. 2020CURL: Contrastive Unsupervised Representations for Reinforcement Learningcited by 760
  11. 2020Reinforcement Learning with Augmented Datacited by 468
  12. 2022DayDreamer: World Models for Physical Robot Learningcited by 400
30 more
  1. 2024Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperationcited by 391
  2. 2021Mastering Visual Continuous Control: Improved Data-Augmented Reinforcement Learningcited by 372
  3. 2019Improving Sample Efficiency in Model-Free Reinforcement Learning from Imagescited by 274
  4. 2023NoMaD: Goal Masked Diffusion Policies for Navigation and Explorationcited by 248
  5. 2016CAD2RL: Real Single-Image Flight without a Single Real Imagecited by 233
  6. 2023Habitat 3.0: A Co-Habitat for Humans, Avatars and Robotscited by 228
  7. 2023ViNT: A Foundation Model for Visual Navigationcited by 222
  8. 2022GNM: A General Navigation Model to Drive Any Robotcited by 171
  9. 2023Reaching the Limit in Autonomous Racing: Optimal Control versus Reinforcement Learningcited by 111
  10. 2016Sim-to-Real Robot Learning from Pixels with Progressive Netscited by 94
  11. 2020Flightmare: A Flexible Quadrotor Simulatorcited by 84
  12. 2020Deep Drone Acrobaticscited by 75
  13. 2021Autonomous Drone Racing with Deep Reinforcement Learningcited by 60
  14. 2018Reinforcement Learning for UAV Attitude Controlcited by 48
  15. 2015From Pixels to Torques: Policy Learning with Deep Dynamical Modelscited by 45
  16. 2019Low Level Control of a Quadrotor with Deep Model-Based Reinforcement Learningcited by 44
  17. 2023Autonomous Drone Racing: A Surveycited by 35
  18. 2018Deep Drone Racing: Learning Agile Flight in Dynamic Environmentscited by 34
  19. 2023Learning to Fly in Secondscited by 31
  20. 2020AlphaPilot: Autonomous Drone Racingcited by 30
  21. 2023Agilicious: Open-Source and Open-Hardware Agile Quadrotor for Vision-Based Flightcited by 30
  22. 2021Time-Optimal Planning for Quadrotor Waypoint Flightcited by 23
  23. 2020Learning to Fly via Deep Model-Based Reinforcement Learningcited by 20
  24. 2023Actor-Critic Model Predictive Control: Differentiable Optimization meets Reinforcement Learning for Agile Flightcited by 19
  25. 2023End-to-end Reinforcement Learning for Time-Optimal Quadcopter Flightcited by 17
  26. 2020Learning Vision-based Reactive Policies for Obstacle Avoidancecited by 12
  27. 2021Human-Piloted Drone Racing: Visual Processing and Controlcited by 9
  28. 2021The Artificial Intelligence behind the winning entry to the 2019 AI Robotic Racing Competitioncited by 8
  29. 2023Sample-Efficient Learning to Solve a Real-World Labyrinth Game Using Data-Augmented Model-Based Reinforcement Learningcited by 8
  30. 2023Time-Optimal Gate-Traversing Planner for Autonomous Drone Racingcited by 7

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 5; 2026: 10.

15 papers.

Sort:

By area:

  1. 2025Accelerating Model-Based Reinforcement Learning with State-Space World Modelscited by 10
  2. 2026SkyJEPA: Learning Long-Horizon World Models for Zero-Shot Sim-to-Real Control of Quadrotorscited by 9
  3. 2025SkyDreamer: Interpretable End-to-End Vision-Based Drone Racing with Model-Based Reinforcement Learningcited by 8
  4. 2025Privileged observations enable rapid and reliable policy discovery directly in the physical worldcited by 3
  5. 2026FlowPilot: Real-Time World-Action Modeling for Agile UAV Navigationcited by 2
  6. 2025All Eyes, no IMU: Learning Flight Attitude from Vision Alonecited by 1
  7. 2025Efficient Model-Based Reinforcement Learning for Robot Control via Online Optimizationcited by 1
  8. 2026Approximate Imitation Learning for Event-based Quadrotor Flight in Cluttered Environmentscited by 1
  9. 2026Generalization of World Models under Environmental Variability for Vision-based Quadrotor Navigationcited by 1
  10. 2026Vision Language Action (VLA) Models for Unmanned Aerial Robotics and Bimanual Manipulation: A Reviewcited by 0
  11. 2026AeroAct: Action-Centered World-Action Models for Language-Conditioned Quadrotor Flightcited by 0
  12. 2026UAVs Meet Embodied Intelligence: Bridging Human Intents and Flying Dynamics Via Harnessing Physical-Digital AI Agentscited by 0
3 more
  1. 2026WOLF: World Model Guided LiDAR Exploration with Predictive Frontierscited by 0
  2. 2026Memory-State Critic for Asymmetric Actor-Critic with Application to Vision-Based Pursuit-Evasioncited by 0
  3. 2026ClearNav: Perception-Aligned Quadrotor Flight Using Differential-Physicscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Curriculum Reinforcement Learning for Quadrotor Racing with Random Obstacles
  2. 2025Mastering Diverse, Unknown, and Cluttered Tracks for Robust Vision-Based Drone Racing
  3. 2021Robust Navigation for Racing Drones based on Imitation Learning and Modularization
  4. 2022Learning Perception-Aware Agile Flight in Cluttered Environments
  5. 2026Bridging Performance and Generalization in Reinforcement Learning for Agile Flight
  6. 2025GRaD-Nav: Efficiently Learning Visual Drone Navigation with Gaussian Radiance Fields and Differentiable Dynamics
  7. 2024Learning Generalizable Policy for Obstacle-Aware Autonomous Drone Racing
  8. 2025FalconGym: A Photorealistic Simulation Framework for Zero-Shot Sim-to-Real Vision-Based Quadrotor Navigation