日本フィジカルAI新聞

世界のフィジカルAIを、日本語で。

宇宙ロボティクスarXiv:2106.00007

DikpolaSatミッション:訓練済み深層ニューラルネットワークによる宇宙機の飛行性能向上と最適制御 — 宇宙物体衝突回避のための軌道制御器

DikpolaSat Mission: Improvement of Space Flight Performance and Optimal Control Using Trained Deep Neural Network -- Trajectory Controller for Space Objects Collision Avoidance

シェア:XThreadsFacebookLINEはてブBluesky

宇宙ミッションDikpolaSatにおいて、訓練済みDNNを軌道制御器として用い、宇宙物体との衝突回避と軌道最適化を図る手法を提案した論文。

著者: Manuel Ntumba, Saurabh Gore, Jean Baptiste Awanyo

分類: cs.RO, cs.LG

原文アブストラクト

This paper introduced the space mission DikpolaSat Mission, how this research fits into the mission, and the importance of having a trained DNN model instead of the usual GN&C functionality. This paper shows how the controller demonstration is carried out by having the spacecraft follow a desired path, specified in the referenced model. Increases can be made by examining the route used to construct a DNN and understanding the effects of various activating functions on system efficiency. The obstacle avoidance algorithm is built into the control features to respond spontaneously using inputs from the neural network for collision avoidance while optimizing the modified trajectory. The action of a neural network to control the adaptive nature of the nonlinear mechanisms in the controller will make the control system capable of handling multiple nonlinear events and also uncertainties that have not been induced in the control algorithm. Multiple algorithms for optimizing flight controls and fuel consumption can be implemented using knowledge of flight dynamics in trajectory and also in the event of obstacle avoidance. This paper also explains how a DNN can learn to control the flight path and make the system more reliable with each launch, thereby improving the chances of predicting collisions of space objects. The data released from this research is used to design more advanced DNN model capable of predicting other orbital events as well.

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. 3GXO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
  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