Distributed State Estimation for Vision-Based Cooperative Slung Load Transportation in GPS-Denied Environments

A research team from Penn State and NASA Langley has published a breakthrough framework enabling teams of drones to cooperatively transport heavy, oversized slung loads without relying on GPS. The paper, presented at the 2026 AIAA SciTech Forum, tackles a critical limitation in prior multilift research, which typically depended on centralized systems or controlled lab setups. The new distributed approach allows scalable payload transport for challenging 'long-tail' cargo without needing to build larger individual aircraft, making operations viable in GPS-denied or sensor-constrained real-world environments.

The technical core is a Distributed and Decentralized Extended Information Filter (DDEIF) that fuses visual data from onboard monocular cameras on each UAV. Each drone detects a fiducial marker on the payload to estimate its relative pose, and the DDEIF architecture robustly combines these measurements across the team, maintaining estimation accuracy even during individual sensor dropouts or communication losses. This payload state estimate then feeds into closed-loop trajectory tracking control. Monte Carlo simulations in Gazebo successfully demonstrated the system's effectiveness, including resilience during communication failures. The work represents a significant step toward practical, resilient multi-agent aerial transportation systems for logistics and disaster response.