# JuggleRL: Mastering Ball Juggling with a Quadrotor via Deep Reinforcement Learning

arXiv: https://arxiv.org/abs/2509.24892 · HTML: https://mlsanity.com/paper/2509.24892/

## Key facts

- Cited by 1 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 15 of its 27 references that are in the map.
- First posted on arXiv 29 Sep 2025.

## Placed in

- Deep reinforcement learning methods / [RL for robot navigation and UAV control](https://mlsanity.com/deeprl/rl-for-robot-navigation-and-uav-control/)

## Builds on (most cited first, 15 of 15)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning](https://mlsanity.com/paper/2108.10470/) (2021)
- [Orbit: A Unified Simulation Framework for Interactive Robot Learning Environments](https://mlsanity.com/paper/2301.04195/) (2023)
- [Control of a Quadrotor with Reinforcement Learning](https://mlsanity.com/paper/1707.05110/) (2017)
- [EGO-Planner: An ESDF-free Gradient-based Local Planner for Quadrotors](https://mlsanity.com/paper/2008.08835/) (2020)
- [UrbanVideo-Bench: Benchmarking Vision-Language Models on Embodied Intelligence with Video Data in Urban Spaces](https://mlsanity.com/paper/2503.06157/) (2025)
- [OmniDrones: An Efficient and Flexible Platform for Reinforcement Learning in Drone Control](https://mlsanity.com/paper/2309.12825/) (2023)
- [CityNavAgent: Aerial Vision-and-Language Navigation with Hierarchical Semantic Planning and Global Memory](https://mlsanity.com/paper/2505.05622/) (2025)
- [FlightGoggles: A Modular Framework for Photorealistic Camera, Exteroceptive Sensor, and Dynamics Simulation](https://mlsanity.com/paper/1905.11377/) (2019)
- [Learning Minimum-Time Flight in Cluttered Environments](https://mlsanity.com/paper/2203.15052/) (2022)
- [EGO-Swarm: A Fully Autonomous and Decentralized Quadrotor Swarm System in Cluttered Environments](https://mlsanity.com/paper/2011.04183/) (2020)
- [What Matters in Learning A Zero-Shot Sim-to-Real RL Policy for Quadrotor Control? A Comprehensive Study](https://mlsanity.com/paper/2412.11764/) (2024)
- [Aggressive Quadrotor Flight Using Curiosity-Driven Reinforcement Learning](https://mlsanity.com/paper/2203.14033/) (2022)
- [Ultra-High-Frequency Harmony: mmWave Radar and Event Camera Orchestrate Accurate Drone Landing](https://mlsanity.com/paper/2502.14992/) (2025)
- [FlightBench: Benchmarking Learning-based Methods for Ego-vision-based Quadrotors Navigation](https://mlsanity.com/paper/2406.05687/) (2024)

## Built on by (most cited first, 1 of 1)

- [RobotWorld: Benchmarking Multimodal Agents for Robot Use Across Diverse Tasks and Embodiments](https://mlsanity.com/paper/2610.10409/) (2026)

## Abstract (arXiv)

Aerial robots interacting with objects must perform precise, contact-rich maneuvers under uncertainty. In this paper, we study the problem of aerial ball juggling using a quadrotor equipped with a racket, a task that demands accurate timing, stable control, and continuous adaptation. We propose JuggleRL, the first reinforcement learning-based system for aerial juggling. It learns closed-loop policies in large-scale simulation using systematic calibration of quadrotor and ball dynamics to reduce the sim-to-real gap. The training incorporates reward shaping to encourage racket-centered hits and sustained juggling, as well as domain randomization over ball position and coefficient of restitution to enhance robustness and transferability. The learned policy outputs mid-level commands executed by a low-level controller and is deployed zero-shot on real hardware, where an enhanced perception module with a lightweight communication protocol reduces delays in high-frequency state estimation and ensures real-time control. Experiments show that JuggleRL achieves an average of 311 hits over 10 consecutive trials in the real world, with a maximum of 462 hits observed, far exceeding a model-based baseline that reaches at most 14 hits with an average of 3.1. Moreover, the policy generalizes to unseen conditions, successfully juggling a lighter 5 g ball with an average of 145.9 hits. This work demonstrates that reinforcement learning can empower aerial robots with robust and stable control in dynamic interaction tasks.