Search papers, labs, and topics across Lattice.
Crazyflow is a unified, GPU-accelerated drone simulator that integrates high-fidelity, differentiable simulations to enhance aerial robotics algorithm development across various methodologies. It significantly outperforms existing simulators, achieving over an order of magnitude faster speeds for single drones and supporting swarm simulations of up to 4000 drones. Notably, Crazyflow enables in-flight reinforcement learning, demonstrated by training a recovery policy in just 0.38 seconds during a physical drone throw, showcasing its potential for real-time applications and rapid algorithm iteration.
In-flight reinforcement learning is now feasible, with Crazyflow training a drone recovery policy in just 0.38 seconds mid-air.
High-quality, large-scale synthetic data from simulations is becoming a cornerstone for pushing the capabilities of robot algorithms. While aerial robotics simulators have evolved to support specialized needs such as fidelity, differentiability, and swarms independently, a unified platform that can synthesize data across all these domains is missing. In this work, we propose Crazyflow, a simulator designed to push the limits of aerial-robotics algorithm development, from model-based to data-driven methods, gradient-based to sampling-based approaches, and single-agent to multi-agent systems. Compared to existing state-of-the-art drone simulators, it achieves speeds more than an order of magnitude faster for a single drone and can simulate thousands of swarms of 4000 drones each. Real-world experiments show Crazyflow supports both analytical-gradient-based policy learning, achieving sub-centimeter trajectory tracking accuracy without domain randomization, and sampling-based obstacle avoidance at speeds exceeding half a billion steps per second. Breaking the traditional train-then-deploy paradigm, we show that its unprecedented speed even enables in-flight reinforcement learning; we demonstrate this by throwing a physical drone into the air and training a recovery policy from scratch in 0.38 seconds, successfully stabilizing the drone. Crazyflow supports multiple levels of simulation abstraction, is directly compatible with all open-source Crazyflie models, and enables rapid reconfiguration across custom drone platforms and applications by providing a light-weight system identification pipeline. By pushing accuracy, speed, and differentiability simultaneously, Crazyflow serves as an open-source resource for synthetic data generation, with emerging capabilities for large-scale parallelization for online, in-execution learning and optimization, opening the door to novel algorithm development.
