Dual-dynamic-absorber switching motion patterns using neural network

2019 Fiscal Year Final Research Report

• Project/Area Number: 17K14615
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Dynamics/Control
• Research Institution: Tottori University
• Principal Investigator: HONGU Junichi 鳥取大学, 工学研究科, 助教 (80781690)
• Project Period (FY): 2017-04-01 – 2020-03-31
• Keywords: 機械力学・制御 / 制振 / 動吸振器 / 自己組織化 / ネットワーク

Outline of Final Research Achievements

To switch the motion of the multi-dynamic absorber, this study focuses on the algorithm of gaits of animals. Developing a dynamic absorber, which can switch its motion pattern by using especially designed nerves system is the final goal of this study in the 2017-2019. First year, we have manufactured a dual-dynamic absorber with two-pendulum and discussed its dynamics. Second year, we have attempted to design the nerves system switching two motion patterns of the dual dynamic absorber. Last year, we have proposed a new network structure to install in the controller of the dual dynamic absorber system. This network structure has two-hubs which correspond to two motion patterns, and utilizes the characteristic of a hierarchical network and an interconnection network. As a results of this study, we have cleared that the hubs in a network has important role to motion pattern changes of the dual-dynamic absorber.

Free Research Field

機械力学

Academic Significance and Societal Importance of the Research Achievements

本研究は，機械装置の知能化を目指し，「自己組織化をいかにして取り込むか」ではなく「自己組織化をいかにデザインするか」にまで焦点を当てるところに意義がある．これにより，自由自在・狙い通りの自己組織化現象をデザインすることが可能になれば，複雑な機構を持つ工作機械・自動車・航空機といった様々な機械への応用が期待でき，本研究の及ぼす工学への影響は大きい．また，これまでに，マルチモード制振には多重動吸振器が有効とされてきたが，受動機構の運動形態の変化に着目した制御手法は存在しない．さらに，外部環境の変化に対して適切な運動形態を選択することで，省エネルギー・高効率なシステムになることが期待できる．