一、課程說明(Course Desc<x>ription)
<br>本課程模組分為三個主要的部分，分別為即時追蹤與地圖建置(SLAM)、基於機器學習之場景理解
<br>(Scene Understanding)與探索導航的動作控制(Action Control)。即時追蹤與地圖建置部分包
<br>含機率模型與相機模型等理論基礎，再搭配2D場景追蹤建圖的實作並介紹RGB-ba<x>sed的3DSLAM。
<br>場景理解的部分包含機器學習的基本概念，再帶到深度學習的技術與目前的物件偵測與語意切割技
<br>術。動作控制的部分則包含路徑規劃與導航演算法，並帶入強化學習的概念來引導行進的路徑。
<br>
<br>二、指定用書(Text Books)
<br>無
<br>
<br>三、參考書籍(References)
<br>Richard S. Sutton and Andrew G. Barto, Reinforcement Learning: An 
<br>Introduction, Second Edition, MIT Press, Cambridge, MA, 2018
<br>Sebastian Thrun, Wolfram Burgard, and Dieter Fox , Probabilistic Robotics, 
<br>2005. (Intelligent Robotics and Autonomous Agents series)
<br>Kevin Murphy, Machine Learning: A Probabilistic Perspective.
<br>Daphne Koller and Nir Friedman, Probabilistic Graphical Models: Principles and 
<br>Techniques, 1st Edition, 2009
<br>Ian Goodfellow, Yoshua Bengio and Aaron Courville: Deep Learning.
<br>
<br>四、教學方式(Teaching Method)
<br>課堂講授、程式作業與專題實作
<br>
<br>五、教學進度(Syllabus)
<br>* Week 1 - Introduction
<br>* Week 2 - Kinematic Model and Path Tracking Control 
<br>    * Control System Basics
<br>    * PID Control
<br>    * Basic Kinematic Model
<br>    * Basic Kinematic Model
<br>    * Differential Drive Vehicle
<br>    * Pure Pursuit Control
<br>    * Kinematic Bicycle Model
<br>* Week 3 - Motion Planning
<br>    * Motion Planning Introduction
<br>    * Path Planning
<br>    * Curve Interpolation
<br>    * Trajectory Planning
<br>    * Path Planning
<br>* Week 4~5: Reinforcement Learning
<br>    * MDP
<br>    * Value Function
<br>    * Bellman Equation
<br>    * Reinforcement Learning
<br>    * Q-Learning / Sarsa / DQN
<br>    * Policy Gradient / Actor-Critic
<br>* Week 6~9 - Lab: Project Environment Building
<br>    * Unity/ROS Environment Building
<br>    * Unity3D(URDF)
<br>    * Solidwork and Unity3D
<br>    * ROSBRIDGE
<br>    * Navigation and Collision Avoidence via RL
<br>* Week 10~11 - SLAM Back-end
<br>    * State Estimation and SLAM Problem
<br>    * Probability Theory and Bayes Filter
<br>    * Kalman Filter / Extended Kalman Filter
<br>    * Particle Filter & Fast SLAM (optional)
<br>    * Graph ba<x>sed Optimization
<br>    * Graph Optimization for 2D SLAM (Bundle Adjustment)
<br>    * Occupancy Grid Map & Laser Beam Model
<br>* Week 12~14 - 3D SLAM
<br>    * Feature Desc<x>riptor 
<br>    * Multi-view Geometry
<br>    * Lie Group & Lie Algebra
<br>    * 3D SLAM: ORB-SLAM
<br>    * Direct Method
<br>    * DNN-ba<x>sed SLAM
<br>* Week 15~16: Project Development
<br>	* Real Scene
<br>	* Simulation Scene
<br>
<br>六、成績考核(Evaluation)
<br>作業: 60% (15% for each HW)
<br>期末專題(含實作、書面報告、口頭報告): 40%
<br>課堂表現: Bonus
<br>
<br>七、可連結之網頁位址 相關網頁(Personal Website)
<br>無
<br>
<br>八、採用AI使用規則 (Indicate which of the following options you use to manage 
<br>student use of the AI)
<br>　 有條件開放，請註明如何使用生成式AI於課程產出 (Conditionally open; please specify 
<br>how generative AI will be used in course output)
<br>