Registered Data

[01011] Analysis and Design of Dynamical Circuits, Systems and Networks

  • Session Time & Room : 5D (Aug.25, 15:30-17:10) @A207
  • Type : Proposal of Minisymposium
  • Abstract : In order to properly design electric circuits and other dynamical systems and networks, it is important to understand their structures and dynamics through mathematical analysis. This minisymposium brings together four engineering researchers working on dynamical circuits, systems and networks to present their recent research results on analysis and design from various perspectives such as discrete harmonic analysis, bifurcation analysis, model order reduction, mathematical programming and contraction mapping principle, with applications to metamaterials, communications, control and signal processing. Through their talks, the importance of mathematics in engineering will be demonstrated.
  • Organizer(s) : Norikazu Takahashi
  • Classification : 94Cxx, 94Dxx, 34Hxx, 65Kxx, 90Cxx
  • Minisymposium Program :
    • 01011 (1/1) : 5D @A207 [Chair: Norikazu Takahashi]
      • [01717] Network dynamics on electric circuits and Maxwell's equations
        • Format : Talk at Waseda University
        • Author(s) :
          • Takashi Hisakado (Kyoto University)
        • Abstract : Electric circuits are the network dynamics of voltages and currents, and electromagnetic phenomena described by Maxwell's equations can be appropriately modeled and designed on the network. In the presentation, we will summarize the properties of electric circuits as discrete harmonic analysis and show that by considering the relationship between Maxwell's equations and electric circuits, a new type of electric circuit can be used to design metamaterials, antennas, etc.
      • [01871] Design of a delayed feedback controller based on bifurcation analysis
        • Format : Talk at Waseda University
        • Author(s) :
          • Koki Yoshida (National Institute of Technology, Toyama College)
          • Keiji Konishi (Osaka Metropolitan University)
          • Naoyuki Hara (Osaka Metropolitan University)
        • Abstract : In addition to stabilizing an operating point, increasing the basin size is required to construct a robust system against disturbances. However, the time delays in the system or control dynamics make it difficult to estimate the basin size analytically. This presentation will show how to design a delayed feedback controller that expands the basin size of the operating point in a DC bus system with the help of a numerical bifurcation analysis tool.
      • [02715] Nonlinear Model Order Reduction for CT Image Reconstruction
        • Format : Talk at Waseda University
        • Author(s) :
          • Yuichi Tanji (Kagawa University)
          • Ken'ichi Fujimoto (Kagawa University)
        • Abstract : X-ray computed tomography apparatus is widely used in clinics to determine whether there is a tumor inside the human body. In this study, we present how to generate a prediction model for image reconstruction using nonlinear model order reduction of the nonlinear dynamical system. For the model reduction, we use proper orthogonal decomposition. We improve the computational techniques so that many observed data can be considered, which provides effective orthogonal vectors for model reduction.
      • [02712] Analysis and Design of Recurrent Neural Networks Generating Desired Sequences of Bipolar Vectors
        • Format : Talk at Waseda University
        • Author(s) :
          • Norikazu Takahashi (Okayama University)
          • Tsuyoshi Yamakawa (Kyushu University)
          • Yasuhiro Minetoma (Kyushu University)
          • Tetsuo Nishi (Kyushu University)
          • Tsuyoshi Migita (Okayama University)
        • Abstract : Recurrent neural networks or RNNs can generate a variety of pattern sequences by the time evolution of each neuron's output. We derive sufficient conditions under which an RNN described by a system of piecewise linear differential equations can generate a desired pattern sequence, and present mathematical programming problems to find its parameter values that satisfy those conditions. Furthermore, we analyze the convergence of state trajectories of an RNN to a limit cycle using a fixed-point theorem and the contraction mapping principle.