文化大學機構典藏 CCUR:Item 987654321/52472
English  |  正體中文  |  简体中文  |  Items with full text/Total items : 46962/50828 (92%)
Visitors : 12455308      Online Users : 528
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version


    Please use this identifier to cite or link to this item: https://irlib.pccu.edu.tw/handle/987654321/52472


    Title: LQR和強化學習倒單擺控制之比較
    Comparison between LQR and Reinforcement Learning on Inverted Pendulum Control
    Authors: 張承熹
    Contributors: 機械工程學系數位機電碩士班
    Keywords: 強化學習
    倒單擺
    獎勵函數
    深度確定性策略梯度
    二次線性最佳化
    控制
    Reinforcement Learning
    Inverted Pendulum
    Matlab
    Reward
    Deep Deterministic Policy Gradient
    Linear Quadratic Regulator
    control
    Date: 2022
    Issue Date: 2023-06-13 10:04:48 (UTC+8)
    Abstract: 隨者電腦軟體的進步,利用電腦軟體模擬機械運動狀態,可更直觀的透過模擬數據去分析機械運動狀態。使用Matlab和Simscape™ Multibody™,除了能夠更方便的計算出如增益值之外,Simscape™ Multibody™利用3D模型,從模型運動過程得知3D模型依照控制器運動的實際狀況。
    LQR和DDPG皆能成功控制倒單擺上甩至直立角度並讓滑車移動至指定位置。切換使用線性二次最佳化控制LQR驅動單擺到目標位置和直立之角度,DDPG則是訓練單擺從下單擺之直接上甩到直立之角度並移動到目標位置。本研究使用Matlab Simscape™ Multibody™來進行分析及3D模型之模擬。二階段LQR和DDPG皆能成功控制倒單擺成功上甩平衡到指定角度,並讓滑車移動到指定位置。本研究亦嘗試並設計一個實體倒單擺機台來驗證模擬的結果。
    With the advancement of computer software, it is possible to simulate the machine motion by using computer software to analyze the machine motion more intuitively through the simulation data. Using Matlab and Simscape™ Multibody™, it is easy to calculate the gain value, and Simscape™ Multibody™ uses the 3D model to know the actual condition of the 3D model movement according to the controller from the model movement process.
    LQR and DDPG can successfully control the inverted pendulum up to an specified angle and move the carriage to the specified position. The switch uses linear quadratic optimization to control the LQR to drive the pendulum to the specified position and the angle, while the DDPG trains the pendulum to move from the lower pendulum directly up to the specified angle and move to specified position. In this paper, Matlab Simscape™ Multibody™ is used for analysis and simulation of 3D model. LQR and DDPG can successfully control the inverted pendulum to balance up to the specified angle and move the carriage to the specified position. In this paper was tried and designed a physical inverted pendulum machine to verify the simulation results.
    Appears in Collections:[Department of Mechanical Engineering ] thesis

    Files in This Item:

    File Description SizeFormat
    index.html0KbHTML128View/Open


    All items in CCUR are protected by copyright, with all rights reserved.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback