Study on a Kernel Function in a Database-Driven Control System

カテゴリ: 部門大会

論文No: SS1-6

グループ名: 【C】2022年電気学会電子・情報・システム部門大会

発行日: 2022/08/24

タイトル(英語): Study on a Kernel Function in a Database-Driven Control System

著者名: Zhu Yuan（広島大学）,Kinoshita Takuya（広島大学）,Yamamoto Toru（広島大学）

著者名(英語): Yuan Zhu (Hiroshima University),Takuya Kinoshita (Hiroshima University),Toru Yamamoto (Hiroshima University)

キーワード: データ駆動型制御|カーネル関数|カーネル密度推定カーネル密度推定|Database Driven Control|Kernel Function|Kernel Density Estimation

要約(日本語): In recent years, developments in control field have been remarkable. As a result, complicated control rules can be applied, such as database-driven control method (DD-PID method), which has been proposed as an effective method for non-linear systems. On the other hand, given the sample data, there is a kernel density estimation method for estimating the density from probability variables. By applying this method to DD-PID method and calculating the similarity between past operation data and its queries, it is possible to automatically determine the number of neighbor data. However, there is no mention of how to select kernel functions. Therefore, this study examines the differences in control performance between different kernel functions, and discusses how to select the kernel function.