Robust Evacuation for Multi-Drone Failure in Drone Light Shows

2026-01-11

arXiv.org

AI Summary

This work proposes a疏散 algorithm integrating trajectory prediction and collision-avoidance strategies to address cascading collisions and safety risks caused by simultaneous failures of multiple drones in aerial light shows. The approach innovatively employs an attention-augmented Social LSTM for the first time to predict the trajectories of failed drones and introduces a dynamic standby drone substitution mechanism to ensure rapid recovery and continuity of the performance. Experimental results demonstrate that the proposed method significantly reduces the probability of surviving drones being struck by falling units, thereby substantially enhancing the overall robustness of multi-drone systems.

Technology Category

Intelligent Robots: Multi-Robot SystemsPlanning, Routing, and Scheduling: Optimization of Spatio-temporal SystemsMultiagent Systems: Distributed Problem Solving

Application Category

Responsible Web: Human-perceived consequences of algorithmic deployment on the webSocial Networks and Social Media: Generative AI / large language models and their impact on social systemsGraph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Drone light shows have emerged as a popular form of entertainment in recent years. However, several high-profile incidents involving large-scale drone failures -- where multiple drones simultaneously fall from the sky -- have raised safety and reliability concerns. To ensure robustness, we propose a drone parking algorithm designed specifically for multiple drone failures in drone light shows, aimed at mitigating the risk of cascading collisions by drone evacuation and enabling rapid recovery from failures by leveraging strategically placed hidden drones. Our algorithm integrates a Social LSTM model with attention mechanisms to predict the trajectories of failing drones and compute near-optimal evacuation paths that minimize the likelihood of surviving drones being hit by fallen drones. In the recovery node, our system deploys hidden drones (operating with their LED lights turned off) to replace failed drones so that the drone light show can continue. Our experiments showed that our approach can greatly increase the robustness of a multi-drone system by leveraging deep learning to predict the trajectories of fallen drones.

Problem

• drone light shows
• multi-drone failure
• evacuation
• robustness
• cascading collisions

Innovation

• multi-drone failure
• trajectory prediction
• evacuation algorithm
• hidden drones
• Social LSTM with attention