2A1-E03 自己組織化ネットワーク素子群における対ノイズ性能の向上

• 出澤 純一

  早大

• 金 天海

  早大

• 尾形 哲也

  京大

• 菅野 重樹

  早大

書誌事項

タイトル別名

• 2A1-E03 Improvement against Noise for Self-Organizing Network Elements(SONE)

抄録

The network learning method named Self-Organizing Network Elements (SONE) was proposed and has been investigated by us. Autonomous exploration of effective output, simple external parameters, and low calculation costs are achieved as functions of a robot system with this method. However, there is the necessity of improving performance against noises for learning more complicated tasks. Therefore, we propose the technique for adjusting the thresholds in Self-Organizing Logic Circuit in this paper. In our experiments, 3Bit operation tasks and controlling Khepera robot, the generation of new elements was appropriately controlled with this technique, and also the network performance against noises is improved.

収録刊行物

• ロボティクス・メカトロニクス講演会講演概要集

  ロボティクス・メカトロニクス講演会講演概要集 2006 (0), _2A1-E03_1-_2A1-E03_4, 2006

  一般社団法人 日本機械学会

キーワード

• Self-Organization
• Neural Network
• Reinforcement Learning
• Topology
• Robot

詳細情報

• CRID: 1390282680913175808
• NII論文ID: 110008694057
• DOI: 10.1299/jsmermd.2006._2a1-e03_1
• ISSN: 24243124
• Web Site: https://www.jstage.jst.go.jp/article/jsmermd/2006/0/2006__2A1-E03_1/_pdf
• 本文言語コード: ja
• データソース種別:
  • JaLC
  • Crossref
  • CiNii Articles
• 抄録ライセンスフラグ: 使用不可