Holiday
DELTA台達106 MaMbMc
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(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。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。
Course keywords: 同步定位建圖(SLAM)、機器學習(Machine Learning)、導航(Navigation)、機器探索(Robotic Exploration)、強化學習(Reinforcement Learning) 一、課程說明(Course Description) 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 (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 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-based的3DSLAM。 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 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。 二、指定用書(Text Books) 無 三、參考書籍(References) Richard S. Sutton and Andrew G. Barto, Reinforcement Learning: An Introduction, Second Edition, MIT Press, Cambridge, MA, 2018 Sebastian Thrun, Wolfram Burgard, and Dieter Fox , Probabilistic Robotics, 2005. (Intelligent Robotics and Autonomous Agents series) Kevin Murphy, Machine Learning: A Probabilistic Perspective. Daphne Koller and Nir Friedman, Probabilistic Graphical Models: Principles and Techniques, 1st Edition, 2009 Ian Goodfellow, Yoshua Bengio and Aaron Courville: Deep Learning. 四、教學方式(Teaching Method) 課堂講授、程式作業與專題實作 五、教學進度(Syllabus) * Week 1 - Introduction * Week 2 - Kinematic Model and Path Tracking Control * Control System Basics * PID Control * Basic Kinematic Model * Basic Kinematic Model * Differential Drive Vehicle * Pure Pursuit Control * Kinematic Bicycle Model * Week 3 - Motion Planning * Motion Planning Introduction * Path Planning * Curve Interpolation * Trajectory Planning * Path Planning * Week 4~5: Reinforcement Learning * MDP * Value Function * Bellman Equation * Reinforcement Learning * Q-Learning / Sarsa / DQN * Policy Gradient / Actor-Critic * Week 6~9 - Lab: Project Environment Building * Unity/ROS Environment Building * Unity3D(URDF) * Solidwork and Unity3D * ROSBRIDGE * Navigation and Collision Avoidence via RL * Week 10~11 - SLAM Back-end * State Estimation and SLAM Problem * Probability Theory and Bayes Filter * Kalman Filter / Extended Kalman Filter * Particle Filter & Fast SLAM (optional) * Graph based Optimization * Graph Optimization for 2D SLAM (Bundle Adjustment) * Occupancy Grid Map & Laser Beam Model * Week 12~14 - 3D SLAM * Feature Descriptor * Multi-view Geometry * Lie Group & Lie Algebra * 3D SLAM: ORB-SLAM * Direct Method * DNN-based SLAM * Week 15~16: Project Development * Real Scene * Simulation Scene 六、成績考核(Evaluation) 作業: 60% (15% for each HW) 期末專題(含實作、書面報告、口頭報告): 40% 課堂表現: Bonus 七、可連結之網頁位址 相關網頁(Personal Website) 無 八、採用AI使用規則 (Indicate which of the following options you use to manage student use of the AI) 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 how generative AI will be used in course output)
MON | TUE | WED | THU | FRI | |
08:00108:50 | |||||
09:00209:50 | |||||
10:10311:00 | |||||
11:10412:00 | |||||
12:10n13:00 | |||||
13:20514:10 | |||||
14:20615:10 | |||||
15:30716:20 | |||||
16:30817:20 | |||||
17:30918:20 | |||||
18:30a19:20 | |||||
19:30b20:20 | |||||
20:30c21:20 |
Average Percentage 92.46
Std. Deviation 6.66
16週課程。
-
-