SailSwarmSwIM enables wind-aware flocking for sailing robot swarms

Autonomous sailing robots face a fundamental problem: their motion is entirely shaped by wind. Unlike drones or ground robots, sailboats cannot directly execute desired speeds or headings—they must respect no-go zones, perform tacking maneuvers, and deal with spatially varying wind that makes some robots fast and others nearly immobile. This creates a fast-slow coordination problem where the flock’s transient heterogeneity can lead to fragmentation. To study this, researchers from multiple institutions developed SailSwarmSwIM, a reduced-order simulator that captures wind-dependent speed and maneuverability, steady and gusty wind fields, and realistic tacking behavior. The simulator provides a controlled testbed for collective behavior algorithms in this unique domain.

Building on the classic Couzin flocking model, the team introduced a speed-weighted social interaction rule. The key insight: instead of treating all neighbors equally, the algorithm increases the social influence of slower robots. This balances two forces—attraction to faster neighbors (which helps maintain movement) and cohesion around slower neighbors (which prevents the flock from splitting apart). Experimental results show that this simple adjustment significantly improves flock polarization and reduces close encounters, making the swarm more robust to wind-driven speed variations. The work, submitted to SAB 2026, offers a modeling framework for adaptive collective behavior in any robotic fleet whose motion capabilities are continuously reshaped by environmental forcing.