Gathering Autonomous Mobile Robots Under the Adversarial Defected View Model

2026-03-06

AI Summary

This study addresses the problem of deterministic finite-time gathering of memoryless autonomous robots in the Euclidean plane under adversarial visibility faults, where up to $K$ robots’ views may be maliciously obstructed. Within the distributed Look-Compute-Move framework, the work proposes two novel algorithms: the first achieves gathering under the fully synchronous (FSYNC) model for the previously open case of $(4,2)$ visibility faults without requiring additional capabilities or global coordinate agreement; the second extends to the more general asynchronous (ASYNC) model, supporting arbitrary $(N,K)$ fault scenarios. Leveraging a non-rigid movement model and a mild assumption of partial axis orientation consistency, both algorithms employ robust observation-computation-movement strategies that guarantee deterministic finite-time gathering under their respective scheduling models.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Computer Vision: Adversarial Attacks & Robustness
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Security and Privacy: Large-scale security measurements
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

This paper studies the gathering problem for a set of $N \ge 2$ autonomous mobile robots operating in the Euclidean plane under the distributed Look-Compute-Move model. We consider oblivious robots executing under the adversarial defected view model, in which an activated robot may observe only a restricted subset of robots due to adversarial visibility faults. Consequently, the information obtained during each Look phase may be incomplete and dynamically altered. The objective is to guarantee deterministic finite-time gathering at a location not known a priori despite such sensing restrictions. We present two distributed algorithms under distinct scheduling assumptions. In the fully synchronous (FSYNC) model, we prove finite-time gathering in the adversarial (4, 2) defected view setting, resolving a previously open case without requiring additional capabilities or coordinate agreement. In the asynchronous (ASYNC) model, we establish finite-time gathering under the general adversarial (N, K) defected view model, where an activated robot observes at most K of the other $N - 1$ robots for any $1 \le K<N - 1$. Both results hold under non-rigid motion. The proposed algorithm for the ASYNC model assumes agreement in the direction and orientation of one coordinate axis.

Problem

• gathering
• autonomous mobile robots
• adversarial defected view
• distributed algorithms
• finite-time convergence

Innovation

• adversarial defected view
• finite-time gathering
• autonomous mobile robots
• distributed algorithm
• asynchronous scheduling