CrazyMARL: Decentralized Direct Motor Control Policies for Cooperative Aerial Transport of Cable-Suspended Payloads
Viktor Lorentz, Khaled Wahba, Sayantan Auddy, Marc Toussaint, Wolfgang Hoenig
AI summary
Problem
Coordinating multiple UAVs to transport cable-suspended payloads is hindered by nonlinear dynamics, external disturbances, and slack-to-taut cable transitions, while prior methods rely on simplified rigid-link assumptions or centralized control that limit scalability and real-world robustness.
Approach
CrazyMARL trains a fully decentralized multi-agent policy using IPPO in GPU-parallelized simulation with heavy domain randomization, mapping local observations directly to 250 Hz PWM motor commands without low-level cascades.
Key results
- 80% recovery rate from harsh conditions versus 44% for classical baselines
- Successful zero-shot sim-to-real deployment on Crazyflie 2.1 hardware
- Stable payload tracking and formation maintenance under 3.5 m/s winds and random disturbances
- Direct PWM control enables safe operation near actuation limits without cascaded controllers
Why it matters
Provides a scalable, communication-free control paradigm for resilient multi-UAV teams operating in unstructured environments for logistics and disaster response.
Abstract
Collaborative transportation of cable-suspended payloads by teams of Unmanned Aerial Vehicles (UAVs) has the potential to enhance payload capacity, adapt to different payload shapes, and provide built-in compliance, making it attractive for applications ranging from disaster relief to precision logistics. However, multi-UAV coordination under disturbances, nonlinear payload dynamics, and slack–taut cable modes remains a challenging control problem. To our knowledge, no prior work has addressed these cable mode transitions in the multi-UAV context, instead relying on simplifying rigid-link assumptions. We propose CrazyMARL, a decentralized Reinforcement Learning (RL) framework for multi-UAV cable-suspended payload trans- port. Simulation results demonstrate that the learned policies can outperform classical decentralized controllers in terms of disturbance rejection and tracking precision, achieving an 80% recovery rate from harsh conditions compared to 44% for the baseline method. We also achieve successful zero-shot sim- to-real transfer and demonstrate that our policies are highly robust under harsh conditions, including wind, random external disturbances, and transitions between slack and taut cable dynamics. This work paves the way for autonomous, resilient UAV teams capable of executing complex payload missions in unstructured environments. Code and videos can be found on the website: https://imrclab.github.io/CrazyMARL. Corresponding author: viktor.lorentz@hhi.fraunhofer.de 1Technische Universit ̈at Berlin, Berlin, Germany. 2Robotics Institute Germany (RIG), Germany. This work was supported by the German Federal Ministry of Research, Technology and Space (BMFTR) under the Robotics Institute Germany (RIG), and the Deutsche Forschungsgemeinschaft (DFG) - 448549715.