# Adversarial Impacts on Autonomous Decentralized Lightweight Swarms

**Authors:** Shaya Wolf, Rafer Cooley, Mike Borowczak  
**arXiv:** [2002.09109](https://arxiv.org/abs/2002.09109)  
**Date:** 2020-02-21  
**Summary page:** https://www.summarizepaper.com/en/paper/2002.09109v1/

## Key points

- Use of swarms in autonomous unmanned vehicles is increasing due to technological advancements
- Swarms exhibit efficient and coordinated behaviors while minimizing computational requirements for each drone
- Susceptibility to external influences is a concern
- Study by Shaya Wolf, Rafer Cooley, and Mike Borowczak explores impacts of adversarial forces on decentralized lightweight swarms
- External influences can compromise swarm functionality, posing risks in scenarios involving human safety or health
- Focus on adversarial swarm using attack vector within movement algorithm of autonomous swarm designed for perimeter surveillance
- Simulations confirmed adversarial swarm could capture significant portions (6-23%) of the perimeter
- Emphasizes importance of understanding and mitigating vulnerabilities in autonomous swarm technologies for safe deployment

## Summary

The use of swarms in autonomous unmanned vehicles has become increasingly prevalent due to advancements in technology. These swarms exhibit efficient and coordinated behaviors while minimizing computational requirements for each individual drone. However, their susceptibility to external influences remains a topic of concern. A recent study by Shaya Wolf, Rafer Cooley, and Mike Borowczak delves into the potential impacts of adversarial forces on autonomous decentralized lightweight swarms. The research investigates how these external influences can compromise the functionality of an autonomous swarm, posing risks in scenarios involving human safety or health. The study focuses on an adversarial swarm that employs an attack vector within the decentralized movement algorithm of a previously defined autonomous swarm designed for perimeter surveillance. Through simulations, it was confirmed that this adversarial swarm could capture significant portions (ranging from 6-23%) of the perimeter, highlighting the potential risks posed by malicious actors seeking to disrupt or manipulate autonomous systems for their own gain. This research emphasizes the importance of understanding and mitigating vulnerabilities in autonomous swarm technologies to ensure their safe and effective deployment in real-world scenarios.