日本フィジカルAI新聞

世界のフィジカルAIを、日本語で。

群制御arXiv:2405.00586

通信制約下での多ロボット探査と静電異常の特性評価戦略

Multi-Robot Strategies for Communication-Constrained Exploration and Electrostatic Anomaly Characterization

シェア:XThreadsFacebookLINEはてブBluesky

火星のような通信インフラのない環境で、複数ロボットが近接時のみ情報共有する分散型と固定基地局型の通信戦略を比較し、静電異常の予測アルゴリズムの有効性をシミュレーションで評価した。

著者: Gjosse Zijlstra, Karen L. Aplin, Edmund R. Hunt

分類: cs.RO

原文アブストラクト

Exploration of extreme or remote environments such as Mars is often recognized as an opportunity for multi-robot systems. However, this poses challenges for maintaining robust inter-robot communication without preexisting infrastructure. It may be that robots can only share information when they are physically in close proximity with each other. At the same time, atmospheric phenomena such as dust devils are poorly understood and characterization of their electrostatic properties is of scientific interest. We perform a comparative analysis of two multi-robot communication strategies: a distributed approach, with pairwise intermittent rendezvous, and a centralized, fixed base station approach. We also introduce and evaluate the effectiveness of an algorithm designed to predict the location and strength of electrostatic anomalies, assuming robot proximity. Using an agent-based simulation, we assess the performance of these strategies in a 2D grid cell representation of a Martian environment. Results indicate that a decentralized rendezvous system consistently outperforms a fixed base station system in terms of exploration speed and in reducing the risk of data loss. We also find that inter-robot data sharing improves performance when trying to predict the location and strength of an electrostatic anomaly. These findings indicate the importance of appropriate communication strategies for efficient multi-robot science missions.

関連論文

  1. 1Boston Dynamics、車輪付きAtlasの研究試作機を紹介研究・技術
  2. 2newmoとKDDIが資本業務提携 2028年に堺市で自動運転タクシー商用化目指す提携・買収
  3. 3NVIDIA、CoRLでGR00T 2とIsaac Lab 3.0を発表新製品・発表
  4. 4GXOとApptronik、Apolloの物流現場試験を拡大 最大1,000台の展開を構想量産・展開
  5. 5Google DeepMind、ロボット向けAIを発表新製品・発表

PR・アフィリエイト広告

OZ GAMING(GPU搭載BTOゲーミングPC)

OZ GAMING(GPU搭載BTOゲーミングPC)の公式広告バナーOZ GAMINGで完成品PCを見る ↗
PR本紙発行元 EmplifAI