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:2503.04942

SAFE-TAXI: 階層型マルチUAS安全自動タキシングフレームワーク — 実行時安全保証とコンフリクト解消

SAFE-TAXI: A Hierarchical Multi-UAS Safe Auto-Taxiing Framework with Runtime Safety Assurance and Conflict Resolution

シェア:XThreadsFacebookLINEはてブBluesky

複数の無人航空機の自動地上走行(タキシング)を、集中型の時空間計画と分散型MPC-CBF制御を組み合わせた階層構造で安全に行うフレームワークを提案し、シミュレーションと小型固定翼機で検証した。

著者: Kartik A. Pant, Li-Yu Lin, Worawis Sribunma, Sabine Brunswicker, James M. Goppert, Inseok Hwang

分類: cs.RO

原文アブストラクト

We present a hierarchical safe auto-taxiing framework to enhance the automated ground operations of multiple unmanned aircraft systems (multi-UAS). The auto-taxiing problem becomes particularly challenging due to (i) unknown disturbances, such as crosswind affecting the aircraft dynamics, (ii) taxiway incursions due to unplanned obstacles, and (iii) spatiotemporal conflicts at the intersections between multiple entry points in the taxiway. To address these issues, we propose a hierarchical framework, i.e., SAFE-TAXI, combining centralized spatiotemporal planning with decentralized MPC-CBF-based control to safely navigate the aircraft through the taxiway while avoiding intersection conflicts and unplanned obstacles (e.g., other aircraft or ground vehicles). Our proposed framework decouples the auto-taxiing problem temporally into conflict resolution and motion planning, respectively. Conflict resolution is handled in a centralized manner by computing conflict-aware reference trajectories for each aircraft. In contrast, safety assurance from unplanned obstacles is handled by an MPC-CBF-based controller implemented in a decentralized manner. We demonstrate the effectiveness of our proposed framework through numerical simulations and experimentally validate it using Night Vapor, a small-scale fixed-wing test platform.

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. 2GXO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
  3. 3newmo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
  4. 4NVIDIA、CoRLでGR00T 2とIsaac Lab 3.0を発表新製品・発表
  5. 5UC BerkeleyとNVIDIAの研究者、AIエージェントでロボット制御コードを仮想環境で改善研究・技術

PR・アフィリエイト広告

FRONTIER(GPU搭載BTOパソコン)

FRONTIER(GPU搭載BTOパソコン)の公式広告バナーFRONTIERでGPU搭載PCを見る ↗
PR本紙発行元 EmplifAI