Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Autonomous Drone Racing: A Survey

Drew Hanover, Antonio Loquercio, Leonard Bauersfeld, Angel Romero, Robert Penicka, Yunlong Song, Giovanni Cioffi, Elia Kaufmann, Davide Scaramuzza

TL;DR

This survey analyzes autonomous drone racing as a rigorous testbed for perception, planning, and control under extreme speeds. It contrasts model-based pipelines with learning-based approaches, highlighting hybrids that combine principled dynamics with data-driven components, and surveys simulators, datasets, and open-source tools that accelerate progress. Key contributions include a comprehensive overview of drone modeling, classical and modern planning/control methods, learning-based perceptual and control strategies, and a forward-looking discussion of safety, multi-agent dynamics, and real-world transfer. The work underscores the shift toward data-driven methods while emphasizing the enduring value of modular, model-informed architectures for reliability and generalization in real-world applications.

Abstract

Over the last decade, the use of autonomous drone systems for surveying, search and rescue, or last-mile delivery has increased exponentially. With the rise of these applications comes the need for highly robust, safety-critical algorithms which can operate drones in complex and uncertain environments. Additionally, flying fast enables drones to cover more ground which in turn increases productivity and further strengthens their use case. One proxy for developing algorithms used in high-speed navigation is the task of autonomous drone racing, where researchers program drones to fly through a sequence of gates and avoid obstacles as quickly as possible using onboard sensors and limited computational power. Speeds and accelerations exceed over 80 kph and 4 g respectively, raising significant challenges across perception, planning, control, and state estimation. To achieve maximum performance, systems require real-time algorithms that are robust to motion blur, high dynamic range, model uncertainties, aerodynamic disturbances, and often unpredictable opponents. This survey covers the progression of autonomous drone racing across model-based and learning-based approaches. We provide an overview of the field, its evolution over the years, and conclude with the biggest challenges and open questions to be faced in the future.

Autonomous Drone Racing: A Survey

TL;DR

This survey analyzes autonomous drone racing as a rigorous testbed for perception, planning, and control under extreme speeds. It contrasts model-based pipelines with learning-based approaches, highlighting hybrids that combine principled dynamics with data-driven components, and surveys simulators, datasets, and open-source tools that accelerate progress. Key contributions include a comprehensive overview of drone modeling, classical and modern planning/control methods, learning-based perceptual and control strategies, and a forward-looking discussion of safety, multi-agent dynamics, and real-world transfer. The work underscores the shift toward data-driven methods while emphasizing the enduring value of modular, model-informed architectures for reliability and generalization in real-world applications.

Abstract

Over the last decade, the use of autonomous drone systems for surveying, search and rescue, or last-mile delivery has increased exponentially. With the rise of these applications comes the need for highly robust, safety-critical algorithms which can operate drones in complex and uncertain environments. Additionally, flying fast enables drones to cover more ground which in turn increases productivity and further strengthens their use case. One proxy for developing algorithms used in high-speed navigation is the task of autonomous drone racing, where researchers program drones to fly through a sequence of gates and avoid obstacles as quickly as possible using onboard sensors and limited computational power. Speeds and accelerations exceed over 80 kph and 4 g respectively, raising significant challenges across perception, planning, control, and state estimation. To achieve maximum performance, systems require real-time algorithms that are robust to motion blur, high dynamic range, model uncertainties, aerodynamic disturbances, and often unpredictable opponents. This survey covers the progression of autonomous drone racing across model-based and learning-based approaches. We provide an overview of the field, its evolution over the years, and conclude with the biggest challenges and open questions to be faced in the future.
Paper Structure (48 sections, 3 equations, 8 figures, 1 table)

This paper contains 48 sections, 3 equations, 8 figures, 1 table.

Table of Contents

  1. Introduction
  2. Why Autonomous Drone Racing?
  3. Task Specification
  4. Drone Modeling
  5. Kinematics
  6. Aerodynamics
  7. Motor and Battery Models
  8. Camera and IMU Modeling
  9. Classical Perception, Planning, and Control Pipeline
  10. Perception
  11. VIO
  12. Additional sensor modalities in VIO
  13. Discussions
  14. Planning
  15. Path planning
  16. ...and 33 more sections

Figures (8)

  • Figure 1: History of drone racing competitions that use real drones for navigating the race track, IROS ADR II photo credit jung2018direct.
  • Figure 2: Architecture 1: A classic architecture for an autonomous system programmed using model-based approaches
  • Figure 3: Top speeds demonstrated on autonomous drones over time from both literature and competition data.
  • Figure 4: Architecture 2: Learned Perception
  • Figure 5: Architecture 3: Learned Planning and Perception
  • ...and 3 more figures